Current stories about AI and making art. Headlines and summaries are written here. The link goes to the source.
Animation World Network
Territory Group, the collective behind Territory Studio and Cantina Creative, has started Lux Meridian to develop features, premium series and animated films with AI-native workflows. The company says it is a production studio rather than an AI tech firm, pairing filmmakers and showrunners with custom generative models built through Adobe’s Firefly Foundry from authorized, proprietary material. It already tested that approach on Brad Peyton’s Dark Zone and launches with a year-long content deal with a major Hollywood studio.
Animation World Network
Added Oct 23, 2026 7:49 pm ET
Animation World Network
Former Pixar story artist Connie He explains how her six-minute short about a sleepless woman imagining the noise upstairs was made at Google DeepMind with a 45-person crew. Storyboards, design and animation were done by hand, and custom Veo and Imagen models fine-tuned on the production’s own hand-painted concept art came in last, turning about 7,000 frames into moving, abstract-expressionist brushstrokes. The film premiered at the Tribeca Festival in June.
Animation World Network
Added Oct 23, 2026 7:49 pm ET
The Stanford Daily
Members of the Stanford band found an AI rendering on the university’s Campus Center project site showing an all-white, all-male group in outfits resembling theirs, and called it inaccurate and offensive “AI slop” when plenty of real photos exist. The image was replaced soon after. A university spokesperson defended AI-assisted imagery as standard for architectural visualization, while band members said some figures looked like real current members, which Stanford’s own AI guidelines prohibit.
The Stanford Daily
Added Oct 23, 2026 7:49 pm ET
UC Berkeley Haas
In a policy article in Science, Abhishek Nagaraj of UC Berkeley Haas, Pamela Samuelson of Berkeley Law and Christopher Jon Sprigman of NYU look at market dilution, the legal theory that floods of AI-made work could harm human creators even without copying. They argue the idea lacks a clear definition, that early studies are hard to read, and that hybrid human-and-AI works make it hard to count. If dilution is shown, they suggest lawmakers may be better placed than courts to deal with it.
UC Berkeley Haas
Added Oct 23, 2026 7:49 pm ET
Geo News
Mity, a computer-generated member of BABYQUEEN, an AI girl group modeled on K-pop acts, announced on Xiaohongshu that she was pregnant and taking maternity leave. The post spread fast, with the controversy reportedly passing 78 million views on Weibo. Citing Sixth Tone, Geo News reports that the creators, students who say they have made no money from the project, admitted the storyline was invented because running several AI personas had become exhausting. Some users criticized them for trivializing pregnancy for attention.
Geo News
Added Oct 23, 2026 3:57 pm ET
BU Today
The Boston Printmakers 2026 North American Print Biennial at Boston University’s 808 Gallery shows 120 prints chosen from more than 2,000 submissions by over 600 artists, running from etching and mezzotint to LEGO relief prints and inkjet ghostprinting. Organizers frame the show around what AI cannot do. “One thing AI doesn’t have is hands,” said exhibit cochair Deborah Cornell, pointing to carved marks that swell, taper and contain mistakes. The free show runs through November 21.
BU Today
Added Oct 23, 2026 3:57 pm ET
Dickinson College
Artifice: Aura and Authenticity from Ancient Molds to Machine Learning opened with a free reception on October 23 at Dickinson College’s Trout Gallery in Carlisle, Pennsylvania. The exhibition starts with cylinder seals from 3500 BCE and moves through Renaissance workshop apprentices and mass production, including a Warhol screenprint, to today’s AI tools. It links old questions about labor, authenticity and the “aura” of an artwork to current worries about generative AI and originality.
Dickinson College
Added Oct 23, 2026 3:57 pm ET
Novi magazin
Digital Art Week Belgrade runs October 27 to 31 at the French Institute in Serbia and Europe House, organized by the Center for Digital Curiosity and Kinopravda Institute. The program opens with Sanja Šikoparija’s solo show In a Grain of Light, includes a one-day Mundane AI workshop on artificial intelligence in everyday creative practice led by Žak Laroš and Uroš Krčadinac, open game-development sessions with Demagog Studio, a lecture by Vladan Joler on the invisible structures of the digital world, and a closing symposium called Protocologies.
Novi magazin
Added Oct 23, 2026 3:57 pm ET
TechNode
ByteDance’s cloud arm Volcano Engine released Seedance 2.5 4K on October 23, pitching it at film and television work. The company says extra training sharpens facial expressions, small movements and fine detail, and a Draft mode lets teams test shots and refine storyboards as cheaper 480p previews before rendering a final 4K version with 10-bit color. For now it is open only to selected enterprise customers, and some production and post-production houses in China and abroad have started testing it.
TechNode
Added Oct 23, 2026 11:53 am ET
Runtime Wire
Image-generation startup Ideogram released Vector 1.0 on October 22. The model creates editable SVG vector artwork from a text prompt or converts an existing image into vectors, which matters for designers who need to recolor, resize and adapt assets after they are generated. Ideogram says it is available in its app, through its API and on partner platforms, with image-to-vector conversion starting at $0.10 per output.
Runtime Wire
Added Oct 23, 2026 11:53 am ET
Global Economic Times
Project A.I.M., opening October 24 at the W Museum in Iksan, South Korea, and running to November 22, marks a decade since Lee Sedol’s match with AlphaGo by framing humans and AI as collaborators rather than rivals. Painters Kim Sang-tae and Lee Min-jung and AI prompt engineer Jeon Seung-hoon show existing and new work beside AI images generated from each artist’s statements, with the prompting process on view. Free sessions on four dates let visitors write prompts and refine images themselves.
Global Economic Times
Added Oct 23, 2026 11:53 am ET
Paolo Cirio
Artist and activist Paolo Cirio presented Nudify AI at the Jeu de Paume in Paris on October 23, a project supported by the art center’s Online Spaces program. Cirio says he collected a sample of 15,000 AI-generated images from Grok.com and used AI detection to find sexualized pictures of people made without consent. The work turns anonymized, stylized fragments of those images into a searchable database and digital collages, publishes no faces or identifying details, and argues for regulation and independent moderation.
Paolo Cirio
Added Oct 23, 2026 11:53 am ET
NEWART Centre
The NEWART Centre in Reus, Spain, opened AND PRAXIS on October 22 with works by more than thirty pioneers of telematic, interactive, generative and virtual reality art, all authors in The Bloomsbury Encyclopedia of New Media Art. There is no curatorial selection: each artist, from Stelarc and Jeffrey Shaw to William Latham and Victoria Vesna, picked the piece they want to be remembered by. The show is also the start of a living archive meant to preserve works whose machines and formats are disappearing.
NEWART Centre
Added Oct 23, 2026 11:53 am ET
LACMA
The Los Angeles County Museum of Art said on October 22 that members of its LACMA Digital Leaders group and other figures in the digital art world will give the museum more than 200 works from their own collections over the next several years. The gift spans pre-internet computer art from the mid-20th century through recent work made with artificial intelligence, and the museum says it will help place digital art within the wider story of art history.
LACMA
Added Oct 23, 2026 7:56 am ET
GEroundtable
Los Angeles AI video platform Nova said on October 22 that it has started production on three original series, the disaster thriller The Last Ferry, the science-fiction story Starsown and the folk horror Nian, with a combined budget above $10 million. Signed actors play the leads, with authorized digital doubles for large effects scenes, while AI is used from script breakdown and storyboards through every shot. The company is aiming at streaming platforms and expects the first title in summer 2027.
GEroundtable
Added Oct 23, 2026 7:55 am ET
Film Daily
Morphic, an AI platform for filmmakers and studios, unveiled BLOOD in London on October 22, its first original series made with full creative control by its own team. Set in Wallachia in 1846, the thriller follows nine lords tied by a secret pact who are killed one by one behind sealed doors. The first episodes are out on Morphic and YouTube ahead of Halloween, with a new one planned each week, and the company says the show is also a test bed for its production tools.
Film Daily
Added Oct 23, 2026 7:55 am ET
Aju Press
Memory for Tomorrow, a youth exhibition at The Star Gallery in Insadong organized by the children’s art academy Sally Mom Art, widened its usual environmental theme to history, culture and the values students want to pass on. AJP spoke with participants including a Toronto high schooler who painted the four seasons, and with the academy’s director, who says that because AI can produce an image in seconds, letting students work through their own trial and error over weeks or months matters more than ever.
Aju Press
Added Oct 23, 2026 7:55 am ET
Amsterdam Dance Event
At Amsterdam Dance Event’s Lab Village in Westerpark, Korean artist Yoon Chung Han presents 23 virtual hybrid babies, each tied to a different chromosomal anomaly and dreaming in its own incubator. Their stories are generated from genetic data in collaboration with AI, and visitors can stir them by speaking into a microphone. Commissioned by the Seoul Foundation for Arts and Culture, the free installation recasts genetic difference as imaginative power rather than deficiency.
Amsterdam Dance Event
Added Oct 23, 2026 7:55 am ET
Mealey’s
A federal magistrate judge in Los Angeles dismissed a copyright lawsuit on October 22 after the digital creator behind the popular two-panel “Running Away Balloon” comic and an AI marketing company resolved their dispute. The suit had alleged that the company made the comic available on its own site without permission. The terms of the resolution were not described in the public summary; the case ends with a dismissal order based on the parties’ stipulation.
Mealey’s
Added Oct 22, 2026 7:55 pm ET
ai&
Japanese AI infrastructure company ai& launched ai& video on October 22, a service that turns text, photos or existing clips into new footage on its own servers in Japan. Creators can describe a scene in Japanese, animate a still photo using its subject and lighting, or extend a clip, and some models add audio. It runs in the browser or by API with pay-as-you-go pricing. Businesses can also train private models on characters and footage they hold the rights to, and the company says uploads for ordinary generation are never used for training.
ai&
Added Oct 22, 2026 7:55 pm ET
Business Standard
ShopOS, which has raised $20 million from Binny Bansal’s 3STATE Ventures, launched Sloosh.ai on October 22. The platform is built on the same engine the company uses to make product images, reels and creator-style clips for brands, which pairs AI agents and reusable workflows with human experts. It is now open to agency teams, filmmakers, photographers, artists and other creators globally, with a connector for AI assistants and an Adobe After Effects plugin, and more integrations planned.
Business Standard
Added Oct 22, 2026 7:55 pm ET
TechCrunch
Canva released Canva Craft on October 22, an image generation model built by its own research team, with help from researchers who came over in the Leonardo.ai acquisition. It works in two passes: it first plans layout, scale, hierarchy and text, then renders images up to 2K in about 30 aspect ratios. Instead of photographs, it was trained on designed templates from Canva’s library such as posters, menus and infographics, so outputs can be turned into layered, editable designs. It is available to all users in Dream Lab and through the Canva API, with faster, cheaper variants promised soon.
TechCrunch
Added Oct 22, 2026 3:50 pm ET
The Verge
Ahead of the October 23 premiere of Wildwood, the stop-motion studio’s most ambitious film, Laika CEO and director Travis Knight told The Verge that tools only have meaning in an artist’s hands, and that AI trained on artists’ work and regurgitated without a person behind it amounts to theft. The film took five years, more than 230 puppets and 49 animators, with over 1,000 artists involved, and it adds new puppet tech such as breathing diaphragms and better musculature. Knight says he hopes AI helps in medicine and science but fears it could be devastating without good stewardship.
The Verge
Added Oct 22, 2026 3:50 pm ET
Frieze
In an essay for Frieze, painter David Salle describes working since 2022 with technologist Grant Davis on a custom model trained deliberately narrowly on his 1999–2001 Pastoral paintings. Its distorted reworkings are printed on linen and become the ground for new canvases, which he sometimes follows and sometimes paints out entirely, comparing the process to psychoanalysis, surrealist automatism and Rauschenberg’s found objects. He argues that because the machine predicts the probable, it cannot replace good artists. His Venice Paintings are at Thaddaeus Ropac Paris Pantin until December 23.
Frieze
Added Oct 22, 2026 3:50 pm ET
Music Business Worldwide
In its answer to Universal and Sony’s second lawsuit, filed October 20 in federal court in Massachusetts, Suno concedes it still holds some of the sound recordings used for training, saying the labels had asked it to preserve training data. The labels argue that Suno’s new v6 models, trained partly on outputs of earlier models, simply launder the original infringement. Suno also admits users have been allowed to name artists in prompts since January 2026, which it says are converted into style descriptions, and it maintains that any copying is fair use.
Music Business Worldwide
Added Oct 22, 2026 11:50 am ET
Music Business Worldwide
Reservoir filed two complaints on October 21 in the Northern District of California. One says Suno copied 1,050 of its compositions, including Both Sides Now and Ring of Fire, to train its music generator, and pleads just over $186 million across its counts; the other says Anthropic trained Claude on the lyrics of 853 of its songs. Both suits ask for training copies, including those in model weights, to be impounded or destroyed, and they cite Suno’s own licensing deals as proof that a market for training licenses already exists.
Music Business Worldwide
Added Oct 22, 2026 11:50 am ET
Billboard
Universal Music Group announced on October 22 a licensing deal with TYGER, a platform that uses AI to adapt music live to a player’s gameplay. It is TYGER’s first agreement with a major label; until now its catalog of more than 400,000 songs came mostly from independent artists. Select UMG songs will now be available to TYGER users, and both companies frame it as a new way for artists to reach gaming audiences, following UMG’s earlier AI remix licensing deal with Spotify.
Billboard
Added Oct 22, 2026 11:50 am ET
Media Arts Cities
The 15th edition of KIKK Festival runs October 22 to 25 in Namur, Belgium, with more than 40 talks, an art trail of over 20 installations across the city, an immersive Playground exhibition at Le Pavillon, workshops for all ages and three nights of performances. This year’s program covers artificial intelligence and creative coding alongside bioart, design, virtual reality and data visualization, with interactive works that invite visitors to experiment.
Media Arts Cities
Added Oct 22, 2026 11:50 am ET
Variety
Speaking at the Busan International Film Festival’s market, James Rustad, global head of GenFX at Netflix’s in-house VFX studio Eyeline, walked through a four-stage workflow of setup, narrative, look and delivery. Generated options are judged against the story before moving into lighting, period detail, compositing and technical checks. He pointed to the series Brazil ’70: The Third Star, with more than 1,300 effects shots, and to a team of 45 artists working in a hybrid of generative and traditional techniques.
Variety
Added Oct 22, 2026 11:50 am ET
PR TIMES
Tokyo company SPEED EFFECT, working as DELINE PRIME, put its feature 星の唄が聞こえる (I Can Hear the Song of the Stars) on YouTube on October 22, Japan’s Anime Day. The studio says three people made it in about three months of active work, generating roughly 1,500 cuts with Seedance and MiniMax, retaking each about five times on average. Images came from ChatGPT image, Nano Banana and Midjourney and music from Suno, while the voices are mostly human actors with some AI voice. The story follows an executive who can only measure himself by results as he finds his way back to feeling.
PR TIMES
Added Oct 22, 2026 7:52 am ET
Folha de S.Paulo
Luiza Mahin, Maria Felipa and Tereza de Benguela, three leaders of resistance in Brazilian history, have long been illustrated with the same generic portrait. The Faces Negras Importam project pairs historians with an art direction team that feeds their findings into AI tools to create an individual image of each woman. Historian Aline Najara based Luiza Mahin’s portrait on an autobiographical letter by her son, the abolitionist Luiz Gama, and says the work is about recognizing the many women whose faces were never recorded.
Folha de S.Paulo
Added Oct 22, 2026 7:52 am ET
Xinhua
A Xinhua report from a film industry park in Suzhou follows Qin Wenshan, a former rail safety officer whose AI short videos drew hundreds of millions of views and who now runs a team where AI agents break down scripts, design characters and build scenes. His team says it finished the sci-fi animated short PandaPawz in 28 days for about 30,000 yuan. Another creator, Li Nan, describes a two-person workflow that starts with a human-written outline and storyboards before AI handles frames, voices and editing, and the piece cites industry figures on sharply shorter production schedules.
Xinhua
Added Oct 22, 2026 7:52 am ET
Romanian Art Week
Opening October 22 at Atelier 030202 during Romanian Art Week and running to November 22, Artificial Beauty is built in three parts. One uses AI to alter, complete or reinvent family photographs, another shows short video interviews with artists, theorists and fashion designers inside a media installation, and the third gathers hand-made work by painter Arina Bican, textile artist Sergiu Chihaia and Mircea Modreanu. The organizers say the show offers no verdict on AI and instead looks at how fashion, bodies and visual standards are shifting.
Romanian Art Week
Added Oct 22, 2026 7:52 am ET
ICK Dans Amsterdam
At ICK Dans Amsterdam during Amsterdam Dance Event, Oostrik’s White Space is a spinning blue-and-white light sculpture with generative music by Marcel Wierckx that responds to visitors’ poses, reaching its climax only when people coordinate their positions. Alongside it, SMILE turns visitors’ portraits into flawless, idealized AI replicas using off-the-shelf image generators with minimal intervention, keeping their biases about race or gender on display as a critique of polished machine illusions. Shows run October 22 to 25, with free entry on the weekend.
ICK Dans Amsterdam
Added Oct 22, 2026 7:52 am ET
Courthouse News Service
At a summary judgment hearing in San Jose on October 21, lawyers for Concord, Universal and other publishers argued that Claude reproduces copyrighted lyrics, sometimes word for word, and that the new songs it writes in an artist’s style compete with human work. Anthropic’s lawyers countered that users are the ones prompting for lyrics, that the model learns from vast amounts of text rather than storing songs, and that the record shows no harm to lyric licensing. Judge Eumi Lee pressed both sides on market harm, a key fair use factor, in a case that could become the first AI copyright suit to go to trial.
Courthouse News Service
Added Oct 21, 2026 7:55 pm ET
The Guardian
Writing for the Guardian on October 21, Blake Montgomery looks at where artists are drawing the line on AI music. Grimes’s new album Psy Opera is described as AI-inspired rather than AI-generated, with her team credited for the work, while A$AP Rocky says he prefers to make music himself and Kelly Clarkson suggests AI acts should have a separate chart. He points to Spotify filtering AI slop from recommendations and to Billboard charts led by voice-and-guitar records as signs that listeners are leaning away from machine-made music.
The Guardian
Added Oct 21, 2026 7:55 pm ET
The Conversation
Published October 21, this essay starts with a Reddit thread mocking a poster for the National Gallery of Victoria’s triennial that uses an image from Ayoung Kim’s Delivery Dancer’s Arc: Inverse, a work about gig-economy riders controlled by an AI system. The author argues that critics are lumping together AI replacing artists and artists choosing AI as a medium to expose its effects on labor. Drawing on Joanna Zylinska and Victor Burgin, it asks whether galleries can show critical AI art without turning that critique into marketing.
The Conversation
Added Oct 21, 2026 7:55 pm ET
Amsterdam Dance Event
Running free day and night in Westerpark during Amsterdam Dance Event week, Charged surrounds each visitor with a cloud of light that follows their movement. Like magnetic charges, some visitors’ clouds push each other away while others merge, so the piece grows calmer or more chaotic depending on the crowd. The visuals are built in TouchDesigner and the tracking uses the open-source YOLO model, which in this work only knows where people are and how they move, with no images stored.
Amsterdam Dance Event
Added Oct 21, 2026 7:55 pm ET
Whitney Museum of American Art
Launched October 20 on whitney.org, the online commission extends Jazsalyn’s 2024 Warpmode card deck, forty prompt cards organized around Afrofuturism, Afronowism, Afrovoidism, and Afropocalypse, into an interactive web piece. The artist frames Ancestral Intelligence as an alternative idea of AI, rooted in knowledge, skills, and rituals passed down in Indigenous and diasporic communities. Visitors answer prompts anonymously, and their responses become part of the work as a visualized memory cloud. It was organized by Christiane Paul, the Whitney’s curator of digital art.
Whitney Museum of American Art
Added Oct 21, 2026 3:55 pm ET
Big Think
In a long essay published October 21, Thomas Moynihan profiles London artists and backers who treat AI as raw material rather than a threat. He describes a robot dog fight hosted by Yaya Labs at the Outernet club, the incubator Restless Egg and its founder William Morgan, Alex Quicho’s Girl Intelligence project, and PC Music artist felicita. His argument is that Britain’s long habit of turning industrial alienation into art, from the Romantics to the CCRU collective, helps explain why its creatives are engaging with AI in weird, critical ways instead of simply rejecting it.
Big Think
Added Oct 21, 2026 3:55 pm ET
ACGCH
The Japanese studio, which mixes hand-drawn animation with 3DCG, had invited spring 2027 graduate animators who could draw what they aimed for with Blender or image-generating AI. After a wide reaction online, it posted a statement on October 17 saying it has not decided to bring image-generating AI into production, does not intend to use AI images as finished animation drawings, and will weigh copyright, training data, and staff impact before adopting anything. The revised listing now emphasizes 3DCG tools such as Blender.
ACGCH
Added Oct 21, 2026 3:55 pm ET
Variety
At the World Soundtrack Awards industry days, held October 8 to 10, the European Composer and Songwriter Alliance said it is lobbying around the EU AI Act for three demands: consent before training on copyrighted music, fair payment, and disclosure of datasets. ECSA president Helienne Lindvall warned musicians to read AI services’ terms before uploading their work. Avatar: Fire and Ash composer Simon Franglen and others worried that generative tools are eating the entry-level jobs where young composers learn the craft.
Variety
Added Oct 21, 2026 3:55 pm ET
PR Newswire
Announced October 21, Agent P lets users describe what they want in plain language, then plans the steps and routes each task to a model from Pixazo’s library, showing which model it used if asked. The company says it handles image edits and character-consistent visuals, multi-scene video, VFX tasks like rotoscoping, relighting, and background replacement, short vertical series with a consistent cast, and soundtracks. It is available on the Pixazo Playground.
PR Newswire
Added Oct 21, 2026 3:55 pm ET
Google
Visionaries opened October 20 at the Petit Palais, timed with Art Basel Paris, and runs through January 30, 2027. Artist Prune Nourry made monumental plaster portraits of 18 overlooked sculptors who worked in Paris in the late 1800s and early 1900s, including Camille Claudel, Jane Poupelet, and Nancy Elizabeth Prophet. Because few photos of their workspaces survive, a 10-minute documentary with director Vincent Lorca uses Google’s generative image tools and a Gemini-built 3D tool to turn archival photos into studios viewers can move through. Google’s blog also describes a sequence where AI editing makes the women students slowly fade out of an old photo of Antoine Bourdelle’s studio, leaving only the men.
Google
Added Oct 21, 2026 11:52 am ET
Paste Magazine
The Dave Meyers-directed clip drew criticism in August when fans spotted “AI artists” in the credits, later relabeled as VFX artists. Singer Grian Chatten told the Los Angeles Times that the main scenes used practical effects and prosthetics from more than 110 people, but someone in post-production added AI to the workflow without telling the band, their label, or management. Paste reports that the band now plans to write a strict no-AI rule into its music video contracts, and that Chatten called the technology terrifying for what it could take from human culture.
Paste Magazine
Added Oct 21, 2026 11:52 am ET
PR Newswire
Announced October 21, the integration lets SoundCloud artists pull their own tracks into Moises for stem separation, chord and lyric detection, AI-assisted mixing and mastering, and real-time collaboration, then post the finished version back to SoundCloud to reach fans and earn money where eligible. Moises also highlights Moi, an assistant that can set up sessions and mix to a brief. In the announcement, both companies say they don’t train on artists’ music without consent, and Moises says its models use only licensed material and don’t generate complete songs.
PR Newswire
Added Oct 21, 2026 11:52 am ET
Visit Stockholm
The group show runs October 21 to November 29 at Fotografiska in Stockholm. Visit Stockholm’s listing says it includes a photographer whose fabricated pictures fooled the world, an artist who lets visitors talk to an AI agent trained on her own life, childhood, and heartbreak, and a project where archives and algorithms mutate into strange new species. The listing’s lead image is credited to Kevin Abosch.
Visit Stockholm
Added Oct 21, 2026 11:52 am ET
The Asia Business Daily
As ___ As Possible opens October 23 at Centre Pompidou Hanwha in Seoul’s Yeouido and runs free through November 13. Anthropic paid for the new works and let the five teams use its Claude model while making them, in its first collaboration with a Korean arts foundation. The show is built around two AI research ideas, interpretability and alignment: Jeong Gahee hangs 103 small monitors showing status codes that never explain why a system is in that state, and Yeom Inhwa pairs a walking platform with augmented reality and an AI guide. The Asia Business Daily reports that Nam Mino, Woobak Studio, and Chu Mirim round out the lineup.
The Asia Business Daily
Added Oct 21, 2026 7:51 am ET
Rolling Stone
The AFM, which represents more than 80,000 musicians, started a public campaign on October 20 ahead of November talks on its Sound Recording Labor Agreement with Universal, Sony, Warner, and Disney’s Hollywood Records. The union wants members to be able to approve or refuse the use of their recordings to train AI models, to get clear credit, and to be paid both for AI use and for licensing deals labels strike with tech companies. Rolling Stone reports the union’s president framing the push around consent, credit, and compensation.
Rolling Stone
Added Oct 21, 2026 7:51 am ET
NBC News
Spotify said on October 20 that it has removed more than 50 million tracks that appear to be AI-generated from its curated recommendations so far in 2026. To find them, it studied six months of songs listeners kept skipping or ignoring, noticed they overlapped heavily with low-effort AI audio, and combined that signal with metadata and outside AI-detection tools to retune its recommendation system. NBC News notes the company admits detection tools have limits, and that more than 300,000 artists covering most of what people search for are now verified on the platform.
NBC News
Added Oct 21, 2026 7:51 am ET
ARTnews
An exhibition that opened this month at Geneva’s Musée d’art et d’histoire shows a dozen 1762 chalk portraits of Empress Maria Theresa’s children by Jean-Étienne Liotard, alongside early results from INTERART, a project with Oxford, the Idiap Research Institute, and the University of Lausanne’s forensic science school. It grew from Oxford professor Catriona Seth’s discovery that a portrait long called Marie-Antoinette actually shows her sister Marie-Caroline. ARTnews explains how the team is testing biometric face matching as one more tool for art historians, next to provenance, inscriptions, and costume.
ARTnews
Added Oct 21, 2026 7:51 am ET
Kill Screen
Colombian artist Juan Covelli’s solo show Cosmograma en espacio latente, on view at Policroma in Medellín since September 17, centers on a wall of clay plates pressed from a 3D scan of a painted cliff at Cerro Azul in the Serranía de La Lindosa. Each plate keeps the real rock’s surface and is hand-painted in slip and oxides with figures from a neural network Covelli trained on the site’s pictographs. Kill Screen notes that the gallery text frames the model as a third reader of the paintings, alongside archaeology and the Ice Age megafauna story, both of which tend to leave out the Indigenous present.
Kill Screen
Added Oct 20, 2026 7:47 pm ET
Hugging Face
Bria released Fibo 2.0 on October 20, folding text-to-image generation and instruction-based editing into a single model for the first time in its FIBO line. It comes as a full Dev model meant for fine-tuning and a faster Edit Turbo variant, with weights on Hugging Face for non-commercial use only; commercial use runs through Bria’s API or a separate license. Bria says it was trained only on licensed data and improves text rendering, texture, and anatomy, and the model card lists it first for image editing among open-weights models on the Artificial Analysis leaderboard as of launch day.
Hugging Face
Added Oct 20, 2026 7:47 pm ET
Coco Labs
Coco Labs introduced Vyb on October 20, pitching it as a design model that respects a user’s own fonts, colors, images, and layouts rather than nudging every design toward the same look. Instead of generating a flat image, Vyb assembles designs layer by layer, so text stays text and graphics stay vector, using a new design language the company built called Cocolang, and it can call image generators when it needs new material. In the company’s own tests of 200 edits, Vyb changed far fewer unrelated pixels than other models when asked for small fixes like a text color change.
Coco Labs
Added Oct 20, 2026 7:47 pm ET
La Nouvelle Tribune
Intelligence artificielle, droits devant!, organized by Cartooning for Peace, is on at the FabLab of the Institut français de Casablanca through October 30 with free admission. The exhibition gathers press cartoons from around the world that use satire and visual metaphor to ask who controls what we see online, how much sway algorithms should have over our choices, and what happens as machines take on work once done by people. La Nouvelle Tribune says the aim is to get visitors, especially younger ones, thinking critically about AI instead of treating it as simply good or bad.
La Nouvelle Tribune
Added Oct 20, 2026 7:47 pm ET
Screen
Artist and filmmaker Vikki Bardot’s A Woman Asleep has been picked for the main competition of Germany’s FilmFestival Cottbus, running November 3–8, which Screen says is believed to be a first for an AI-generated feature at an established international festival. The semi-autobiographical film is cut from about 50,000 shots made with AI models Bardot trained and customized herself, narrated by a deliberately flat AI clone of her own voice, with an AI-generated score. It is produced by Turkey’s Spongeworthy Studio and premieres first at the Antalya Golden Orange Film Festival.
Screen
Added Oct 20, 2026 3:55 pm ET
KAIST News Center
A team led by Professor Juhan Nam at KAIST brought two systems to the Jazz AI Lounge stage at this month’s Jarasum Jazz Festival in Korea: an AI pianist called Scott that follows a live performance and takes over the piano part, sometimes adding its own solos, and Deepocket, which listens to a pianist’s improvisation and plays bass and drums in real time. Twelve jazz artists played 26 pieces with the AI across 12 sets, with Scott’s notes played on a Yamaha Disklavier and shown as an animated pianist in augmented reality. The team says the goal is an AI that acts as a bandmate rather than a playback tool.
KAIST News Center
Added Oct 20, 2026 3:55 pm ET
MIT Media Lab
From October 22 to 24 at Symphony Hall, the BSO and conductor Kent Nagano give the world premiere of Tod Machover’s already and not yet, a commission marking the MIT Media Lab’s 40th anniversary. Live AI built by Media Lab PhD student Manaswi Mishra works with each player to create links across the orchestra, which Machover describes as a hopeful picture of shared creative life. The run also includes a pre-concert talk with AI-music pioneer George Lewis and an October 25 program of panels at MIT featuring Holly Herndon on AI as musical practice.
MIT Media Lab
Added Oct 20, 2026 3:55 pm ET
aurèce vettier
The artist aurèce vettier opens off-duty seraphim at DARMO in Paris on October 20, running until November 5. He trains custom AI models on his personal archive, childhood photos, and dreams, then hands the resulting images to traditional craft workshops. The show includes a palladium-gilded bronze cabinet-sculpture, oil paintings, Aubusson tapestries made with painter Vincent Gicquel, and a large group of pietra dura hardstone inlays, all built around guardian figures that have lost their purpose in a world where the rich might defeat death.
aurèce vettier
Added Oct 20, 2026 3:55 pm ET
NoCut News
Korean director Hong Jeong-min’s animated short “Stitching Hearts” follows an elderly man who loses a button tied to memories of someone he loved and finds comfort in a button from his grandchild. Hong wrote the script himself and built the film with Seedance 2.0 for video and Suno for music, finishing in three days with about ten hours of hands-on work. It made the top ten of the short film competition at the ILLUMINATE Film Festival in Sedona, which drew 648 entries from 67 countries. Hong, who has worked on Space Sweepers and Dota: Dragon’s Blood, says what matters is the emotion a short film carries, not how fast AI makes it.
NoCut News
Added Oct 20, 2026 11:57 am ET
Variety
Spotify has added independent music company Believe, including its TuneCore distribution arm, to the licensing deals behind a still-unnamed AI tool that will let Premium subscribers remix or cover songs for an extra fee. Universal, Merlin, and Kobalt signed on earlier, and there is no release date yet. Believe says its artists opt in twice: first to AI services at all, then to Spotify using their work for this tool. Spotify also said it has pulled 50 million low-quality AI tracks out of listener recommendations this year.
Variety
Added Oct 20, 2026 11:57 am ET
Music Ally
Berlin firm Just 4 Noise has released Just 4 Kicks, an AI sample generator that makes fresh kick drums so producers don’t have to dig through the same sample packs as everyone else. The company says it trained the model only on audio it owns, recorded and curated in studios around the world over several months. It sells for a one-time €79 with free updates rather than a subscription, and snare, hi-hat, and full-drum versions are planned.
Music Ally
Added Oct 20, 2026 11:57 am ET
Vertical Slice
The new publishing label from the studio behind Peak and Another Crab’s Treasure told Vertical Slice it will reject pitches that use generative AI, including AI concept art meant only to illustrate an idea. Head of publishing Corey Warning said he would much rather see a developer’s rough sketches, and marketing lead Paige Wilson said the technology is already taking work from indie artists. The label is writing a document that spells out its stance and plans to publish it.
Vertical Slice
Added Oct 20, 2026 11:57 am ET
The Society of Authors
The UK’s Society of Authors and Creators’ Rights Alliance joined the Authors Guild, the Australian Society of Authors, the Writers’ Union of Canada, and the New Zealand Society of Authors to publish a Charter of Ethical AI Use in the Creative Industries. Publishers, agents, prize bodies, and other organisations can sign to pledge transparent and responsible use of generative AI, built on UNESCO’s principles for responsible AI and the CLEAR framework for protecting creators’ livelihoods. Kazuo Ishiguro and Tracy Chevalier are among the authors backing it.
The Society of Authors
Added Oct 20, 2026 11:57 am ET
Vulture
Vulture looks at what happened after Romain Gavras’s music video “Storm” for Gener8ion, with Yung Lean and 66 real schoolboy dancers filmed in one take on 16mm, became a template for AI copies. Higgsfield’s Genjutsu tool, released in late August, can swap out an entire crowd while keeping the original motion and camerawork, and September brought a flood of versions starring pets, celebrities, and worse. Gavras says he turned down AI licensing offers, and a copyright lawyer quoted in the piece says the choreography is likely protected but suing thousands of anonymous users is close to impossible. Gavras answered by restaging the dance live at the VMAs.
Vulture
Added Oct 20, 2026 7:47 am ET
Yonhap News Agency
The Seoul Museum of Art is opening “Lynn Hershman Leeson: First Person Plural,” about 200 works from the American media artist’s six-decade career, at its Seosomun branch through February 21. Hershman Leeson has long tested how identity shifts through video, the internet, AI, and genetic engineering. The show includes her Roberta Breitmore alter-ego project, the interactive “Lorna,” and “Agent Ruby,” an early web chatbot whose expression changes as visitors type to it. Several interactive pieces are shown in Korean for the first time, alongside screenings of her films.
Yonhap News Agency
Added Oct 20, 2026 7:47 am ET
GamesRadar+
Manabu Otsuka, head of the studio behind Jujutsu Kaisen, Chainsaw Man, and Attack on Titan, told Anime Corner he doesn’t rule out AI but treats it like digital tools or 3D: useful at most for making workflows more efficient. He said quality that moves audiences emotionally can only come from people, and that the studio will keep weighing which tools to use and how much. GamesRadar places the comments in a wider pushback from anime and game leaders on generative AI.
GamesRadar+
Added Oct 20, 2026 7:47 am ET
IT Russia
Russia’s largest animation studio used neural networks to help build characters and settings for an adaptation of Yulia Ivanova’s books, The Secret of Charovody. Board chair Yuliana Slashcheva said artists first designed characters and scenes by hand and then used AI to visualize them, finishing in about four months work she estimated would otherwise take at least three years. The studio has previously used AI to restore and upscale classic cartoons, and says it still treats the animator as the creative lead.
IT Russia
Added Oct 20, 2026 7:47 am ET
Music Business Worldwide
In their copyright case against AI music company Suno, Universal Music Group and Sony Music are asking a Boston federal court to make Lightspeed Venture Partners, which led Suno’s Series B, hand over documents and sit for a deposition. They point to a Lightspeed document that discusses the legal risk of training data and label lawsuits. The labels narrowed their subpoena from 20 topics to 16, including Lightspeed’s seat on Suno’s board. Lightspeed has moved to quash the subpoena and can reply by October 20.
Music Business Worldwide
Added Oct 20, 2026 7:47 am ET
VentureBeat
fal released FLUX 3 Max, a version of Black Forest Labs’ FLUX 3 Video that fal post-trained for closer prompt-following and much faster generation. It handles text-to-video, image-to-video with an optional end frame, and a keyframe mode that takes up to 10 images, producing 5 to 20 seconds at 720p with optional 1080p upscaling and optional audio. fal says a 10-second clip takes roughly 12 seconds versus several minutes for Seedance 2.5; those are vendor timings, and VentureBeat notes the model is statistically tied with the top image-to-video leaders rather than clearly ahead. Launch pricing is discounted 75% until November 7.
VentureBeat
Added Oct 19, 2026 7:54 pm ET
Deadline
Former UTA film agent Dante DeSario and media investor Michael Lambert have started Collage, a Los Angeles management company for AI-native production, backed by a seed round from Black Bear and Fundamental Films among others. Its launch roster includes Asteria, real-time 3D animation studio Defect, Berlin’s Lipstick, Verena Puhm’s Matriarch, Critterz creator Native Foreign, and Secret Level. Collage plans to match these studios with studios, streamers, and brands, help price and scope projects, and oversee quality, legal, and delivery work, while investing selectively. Early reader comments on the announcement were sharply split.
Deadline
Added Oct 19, 2026 7:54 pm ET
Lovart
Lovart’s 2.0 update centers on Arti, an assistant that keeps editable memory of your background, collaboration style, aesthetic preferences, and most-used assets, and applies it to new projects. A new design delegation feature lets you hand over a whole brief as a one-off or a daily, weekly, or monthly job with a credit cap and a test run first. Background tools gather references into mood-board reports, flag canvas frames that break a project’s style, and turn reviewers’ anchored comments into a revision plan you approve. Lovart says final creative decisions stay with the designer.
Lovart
Added Oct 19, 2026 7:54 pm ET
GamesBeat
Ludo.ai released the Ludo CLI, a command-line way to use its game-asset generators without a browser. Developers, or coding agents working with them, can create sprites, animations, textures, 3D models, video, music, sound effects, and voices, chain steps like making and then animating a character, run batches from scripts, and save files straight into project folders with preset engine and style settings. It also previews credit costs and supports spending limits. The CLI runs on the same API as Ludo’s existing MCP and REST integrations and is available on Pro and Studio plans via npm.
GamesBeat
Added Oct 19, 2026 7:54 pm ET
Art Magazine AI
Who Speaks The Machine? at KU Leuven’s BAC ART LAB in Belgium runs through October 24 with work by Alex Verhaest, Constant Dullaart, Johanna Reich, Patrick Tresset, and Roel Heremans. Pieces range from Tresset’s robots sketching visitors’ portraits to Reich’s canvases paired with AI-written verse and Dullaart’s paintings referencing the ImageNet dataset. Curated by Michiel Willems, the exhibition is also a research project: visitors can fill in questionnaires about how they perceive AI and digital art, which takes about an hour. Entry is free on selected days.
Art Magazine AI
Added Oct 19, 2026 7:54 pm ET
RTÉ
Researchers in the University of Essex psychology department ran three experiments with more than 650 people in the UK and US, comparing how creators felt about work they made with AI help versus on their own. Participants rated their AI-assisted pieces as less authentic and less creative, felt fewer positive emotions about them, and reported small but measurable rises in disgust, guilt, and shame. Lead researcher Jessica Carson says people also felt they had put less of their “love” into the AI-assisted work. The findings are self-reported and come from a preprint that has not yet been peer reviewed.
RTÉ
Added Oct 19, 2026 3:48 pm ET
Scienmag
Two researchers at Yili Normal University describe GECST in the journal Discover Artificial Intelligence: a GAN-based system for digitally restoring damaged Buddhist murals from China’s Mogao Grottoes. Instead of only filling in missing pixels, it penalizes reconstructions that break the murals’ classical rules for symmetry, spatial hierarchy, and motif balance, then corrects faded color within a historically plausible Dunhuang palette. The goal is restorations that make sense to art historians, not just images that look smooth; the work is digital and does not touch the physical walls.
Scienmag
Added Oct 19, 2026 3:48 pm ET
Hugging Face
Trainer lodestones posted Kromoss on Hugging Face, a full fine-tune of the 5B-parameter MOSS-Image-Base aimed mainly at fixing that model’s blurry output. It drops into existing MOSS-Image workflows in Diffusers and ComfyUI, ships an int8 version for smaller GPUs, and includes an 8-step LoRA for quick drafts alongside the usual 50-step render. The weights are Apache 2.0, so commercial use is allowed. The author’s own prompt-matching scores are modest and were measured on an earlier checkpoint, so it is worth judging the sharpness yourself.
Hugging Face
Added Oct 19, 2026 3:48 pm ET
Hungary Today
Living Patterns: Masters and Students at MOME opens October 20 at the György Ráth Villa of Budapest’s Museum of Applied Arts and runs through January 3, 2027. Curator Judit Horváth says the idea grew out of asking which human factors technology cannot replace in the age of AI, so the show pairs nine teachers from the Moholy-Nagy University of Art and Design with a mentor and a student each, focusing on how skills are passed on rather than on individual portfolios. A catalog of interviews and curator-led tours accompany it.
Hungary Today
Added Oct 19, 2026 3:48 pm ET
Viggle
On October 19, 2026, Viggle introduced JST-2, a world foundation model that represents scenes as persistent 3D “atoms” instead of only pixels, then decodes them as Gaussian splats you can explore in the browser. The company says a scene generates in about a second for under a cent, characters come out animation-ready without a separate rigging step, and full-body motion can be driven from text or video. Scene collisions still use rule-based physics for now; Viggle says learned object interactions are planned for JST-3. A free trial is available in its PINOC browser tool.
Viggle
Added Oct 19, 2026 12:15 pm ET
Türkiye Today
The Ayvalık National Artificial Intelligence Film Festival ran October 16–17 at the Aren Art Workshop in western Türkiye, showing 65 AI-assisted shorts and music videos chosen for artistic merit rather than social-media reach. Festival director Tarkan Kaynar told Türkiye Today that AI helps independents stage scenes they could never afford to shoot, while still warning that directors with original ideas remain essential as the tools reach cinematography, lighting, and editing. Organizers describe this year as a “film days” preview and say they want a weeklong festival in 2027.
Türkiye Today
Added Oct 19, 2026 12:15 pm ET
Google
Google Labs says Spaces in Flow Music can turn a natural-language description into a custom synth or effect, then export it as a VST3 or AU plugin for DAWs such as Ableton Live, Logic Pro, and FL Studio. Launch example “No Chaser,” from producer Khris Riddick-Tynes, is meant to keep vocals and instrumentals studio-sharp wherever they were recorded. For now the plugins are macOS-only, and instrument exports are limited to standard synthetic audio—no sample banks or generative AI audio yet.
Google
Added Oct 19, 2026 12:15 pm ET
Comfy Blog
Comfy Org released custom NVIDIA VFX nodes so RTX GPUs can upscale video up to 4× (or to 8192 px), insert frames from 2× to 16×, and convert SDR footage to BT.2020 HDR with HLG or PQ output—all locally, with the models bundled in the pack. The recommended chain is Super Resolution, then Frame Generation, then HDR, and the same release improves half-float EXR save/load and color-space conversion for VFX pipelines. Frame generation needs an RTX 40-series GPU or newer; install is through ComfyUI Manager.
Comfy Blog
Added Oct 19, 2026 12:15 pm ET
The Guardian
The International Confederation of Music Publishers told Guardian Australia that AI systems have already learned from Australian artists without payment, pointing to Gemini reproducing lyrics by Kylie Minogue, G-Flip, and Nick Cave and to a reported 12-million-song training set pulled from YouTube. The group says Australia is the only one of 62 countries it has tracked that is seriously considering weakening copyright for AI training, after a parliamentary committee asked the prime minister to re-examine the rules while keeping creators in control and paid.
The Guardian
Added Oct 19, 2026 7:54 am ET
AI Affairs
San Francisco start-up VideoGen announced Model 1 on October 19, describing a planning layer that writes the screenplay, shots, and pacing before generating a multi-shot video of up to ten minutes with consistent characters and locations. Its CEO claims quality close to ByteDance’s Seedance 2.5 at less than half the price, but there are no public benchmarks or prices yet and AI Affairs could not test it. A partner-only rollout starts now, alongside applications for a $1 million pool of free credits for filmmakers and students.
AI Affairs
Added Oct 19, 2026 7:54 am ET
Anime News Network
The studio behind work on MF Ghost and Aharen-san wa Hakarenai changed its new-graduate animator listing, which had asked for people who could make illustrations with Blender or image generative AI. The new wording focuses on Blender and 3D tools and says the studio is only studying how AI might support its creators. In a statement, Felix Film said it has not decided to use generative AI for images and would weigh rights issues carefully before adopting any tool.
Anime News Network
Added Oct 19, 2026 7:54 am ET
CBC News
Final-year students at Fanshawe College in London, Ontario told CBC they can spot AI-generated posters by their crowded layouts, missing focal points, and odd fonts, and that brands relying on them start to look untrustworthy. Their program coordinator calls the trend a sea of sameness and expects ordinary viewers to notice it soon. The college still encourages occasional AI use, mostly for research and feedback, with students describing it as a tool rather than a competitor.
CBC News
Added Oct 19, 2026 7:54 am ET
Ministry of Law Singapore
Singapore’s Ministry of Law and its IP office are taking public feedback until 5 PM on October 22 on how AI fits the country’s IP laws. On the copyright side, they ask whether the existing data-analysis exception for AI training is clear enough, how blame should be split between developers, deployers, and users when an output infringes, and how human contribution should be recognized and evidenced when generative AI is used as a tool. Submissions go through FormSG or email.
Ministry of Law Singapore
Added Oct 18, 2026 7:52 pm ET
Anime News Network
The Japan Actors Union released the first version of 21-page guidelines on using generative AI with actors’ and voice actors’ recordings, covering written consent, advance notice, credit and removal requests, compensation for training and use, and special provisions for anime and dubbing. At the press conference, Masako Nozawa, the voice of Goku, said voices are being used without consent, and Bin Shimada described his anger at hearing an AI cover sung in his Broly voice.
Anime News Network
Added Oct 18, 2026 7:52 pm ET
GMA Entertainment
Filipino action director Vincent Soberano won the festival’s first Bold Vision Award and its Gold Award for Best AI Major Motion Picture with Secret Apartment, a 60-minute psychological drama set in Taipei, picked from more than 1,000 entries. He says it began as an experiment in Higgsfield’s Cinema Studio and grew from miniseries episodes into a feature, made in under three months with tools including Seedance 2.0 and Nano Banana Pro. Keeping faces and wardrobe consistent across clips was the hardest part.
GMA Entertainment
Added Oct 18, 2026 7:52 pm ET
TestingCatalog
TestingCatalog reports that references to aurora-red-alpha, aurora-green-alpha, and aurora-blue-alpha appeared in OpenAI’s API this weekend and were quickly pulled, then showed up generating images in blind arena comparisons. Early testers found the red version faster and the green one slower but better looking. Observers guess this could be a next GPT-Image model, but OpenAI has not announced anything, so treat it as a sighting rather than a launch.
TestingCatalog
Added Oct 18, 2026 7:52 pm ET
GitHub (VAST-AI-Research)
VAST, the team behind Tripo, has released code and 5B and 14B weights for EdenLive. A game engine or coding agent runs the rules and outputs a plain greybox view, and EdenLive repaints it live in whatever look a first frame and text prompt set, keeping the layout, camera, and motion. VAST says the 5B model reaches 720p at 24 frames per second or more on a single RTX 5090. The code is Apache 2.0, though some weight components carry their own terms.
GitHub (VAST-AI-Research)
Added Oct 18, 2026 3:53 pm ET
GameSpot
Two days after Total War: Shogun 2 – Complete Edition launched on October 13, Sega and Creative Assembly removed it from sale and reinstated the original 2011 game. Players had posted unit cards with extra fingers and faces melting into helmets, alongside new and old bugs, and the remaster sank to an Overwhelmingly Negative Steam rating. Sega apologized and promised fixes, though its statement did not mention AI; existing owners can keep playing.
GameSpot
Added Oct 18, 2026 3:53 pm ET
KVUE
The second Austin AI Film Festival ran October 16 and 17 at the Zach Theatre, with a day of panels and master classes followed by screenings and an awards night. Founder Aaron Sanchez pitches generative tools as a way for independent filmmakers to build worlds that once took big CGI budgets, while insisting every qualifying film’s story be written by a human. The lineup mixed mostly-AI films with hybrid live-action work.
KVUE
Added Oct 18, 2026 3:53 pm ET
EURweb
In an interview with AllHipHop, Wyclef Jean said AI could cut the huge costs that have kept a big-screen story of Toussaint Louverture from happening, pointing to Danny Glover’s long-stalled Toussaint epic, which never found its financing. EURweb notes he is not alone: director Bill Duke has also described AI as a way to make a long-unfunded film about Black Wall Street without waiting on studio approval.
EURweb
Added Oct 18, 2026 3:53 pm ET
AI Film Festival Japan
The second AI Film Festival Japan runs October 31 to November 2, 2026 at Tokyo Innovation Base, now stretched to three days. Organizers chose 68 official selections and 21 finalists from more than 730 entries across 60-plus countries. New this year is a film-score category where entrants set AI-generated music to an official festival clip, and partners CapCut and ElevenLabs each present a sponsored award, with Alibaba Cloud also on board.
AI Film Festival Japan
Added Oct 18, 2026 11:49 am ET
Digital Music News
In its 2026 Notorious Markets submission to the US Trade Representative, the RIAA named Lalals, a UAE-based site offering voice cloning, AI covers, stem splitting, and song generation, alongside Anna’s Archive and long-listed stream-rippers. Digital Music News notes the site sells soundalike vocal models of well-known artists on plans of $12.99 and $24.99 a month. The filing is the industry group’s request to the government, not a legal ruling.
Digital Music News
Added Oct 18, 2026 11:49 am ET
WIRED
More than two dozen anonymous workers told WIRED that HarperCollins, Simon & Schuster, and Hachette use tools like Claude and ChatGPT to write back-cover and publicity copy, email literary agents, and produce cover art and marketing pieces, often without telling authors. Simon & Schuster staff circulated an open letter against a trial of the Skan AI workflow-monitoring tool, and the CEO said no decision has been made. Hachette says it backs AI for operations but not for creative work.
WIRED
Added Oct 18, 2026 11:49 am ET
NAVA
With the Australian Government weighing copyright changes for AI, the National Association for the Visual Arts welcomed the Prime Minister’s pledge to protect creatives but said that promise needs teeth. NAVA wants AI companies, not individual artists, to carry the burden of getting permission and paying for work they use, and it flags fake Indigenous art made with AI as a growing harm. Its 2025 survey found over 80% of visual artists saw AI as a risk to their income and rights, and it is asking artists to write to their MPs using a letter template.
NAVA
Added Oct 18, 2026 7:55 am ET
UK Music
Three UK parliamentary groups on music, intellectual property, and visual arts met jointly on October 14, 2026 to discuss AI and copyright. Visual artist Matty Burnham said thousands of AI replicas of his artwork are being resold under other names, and folk singer Emily Portman described finding an AI-generated album uploaded to her artist profile without her knowledge. Baroness Keeley also briefed attendees on a House of Lords committee report about unregulated AI.
UK Music
Added Oct 18, 2026 7:55 am ET
ComfyUI Wiki
Released October 16, 2026, the Plenio Music Production System lets a local language model write lyrics, then has YuE2 or MiniMax Music 3 perform them, all inside ComfyUI with no cloud service or API key. It handles stems, mastering to streaming loudness, and export with tags and cover art, and version 1.0 adds a browser player for sharing finished songs. It replaces the same author’s earlier MiniMax music toolkit, which can stay installed alongside it.
ComfyUI Wiki
Added Oct 18, 2026 7:55 am ET
KVR Audio
Initial Audio’s standalone MIDI tool for Windows and macOS now runs a model trained from scratch, aimed at more natural phrasing, better chord movement, and longer musical structure while keeping the feel of your input. The update adds a model picker, personal models shaped by your ratings, session history, and five new prompt packs, and everything runs on your own computer. The free tier rises to 80 generations a month, and Pro stays a $69 one-time purchase.
KVR Audio
Added Oct 18, 2026 7:54 am ET
elseif
At Brazil Game Show 2026, Tripo AI demonstrated Tripo P2.0, which turns text prompts into 3D models built from four-sided polygons so they are easier to edit and animate, plus a Mesh Edit tool that regenerates one selected region without touching the rest. It also showed LEBOT, an arcade fighting game built on its API that turns a player’s photo into a 3D fighter. The report notes generated models may still need cleanup for real production.
elseif
Added Oct 18, 2026 7:54 am ET
Los Angeles Times
The Los Angeles Times reports that AI production companies such as Promise, TrueShort, Native Foreign, and PhantomX are setting up near Sony and Amazon MGM in Culver City, drawn by the studio legacy and cheaper office space. During LA Tech Week, the Culver Steps plaza hosted a job fair for AI studios. TrueShort says a 10-minute fully AI-generated episode costs about $10,000 and takes roughly 10 days, while unions and many Hollywood workers remain worried about lost jobs.
Los Angeles Times
Added Oct 17, 2026 7:56 pm ET
Law Commentary
In an October 14, 2026 filing in Manhattan federal court, the two labels asked a judge to compel DistroKid to hand over monthly counts of AI-generated uploads and details of its AI-detection tools, saying the numbers bear on market harm in their copyright case against Suno. DistroKid had resisted a September subpoena. No ruling has been made yet.
Law Commentary
Added Oct 17, 2026 7:55 pm ET
RIXC
Artists Rasa Smite and Raitis Smits feed herbarium images of wild Andean lupines into AI models to produce speculative plant forms, then turn them into interactive 3D virtual plants. The show opens at RIXC Gallery in Riga on October 23, 2026 as part of the RIXC festival, with free entry through December 12.
RIXC
Added Oct 17, 2026 7:55 pm ET
GitHub (felix-ab)
An independent open-source port runs Black Forest Labs’ new Klein image model on Macs through Apple’s MLX framework, including edits with up to 10 reference images and bounding-box layout. It ships 8-bit and 6-bit builds, with the 6-bit one aimed at 16 GB machines, and the author reports about two minutes per 1024-pixel image on an M4 Pro. A 4-bit build was left out on purpose because it added visible grain.
GitHub (felix-ab)
Added Oct 17, 2026 7:55 pm ET
Joan Mitchell Foundation
In a statement dated October 15, 2026, the foundation that looks after the late painter’s legacy said it will not license Joan Mitchell’s artworks, name, or image for training, testing, or improving AI models, and will turn down any request to animate her paintings or photos or to generate new images or audio of her. It also filed a formal text-and-data-mining opt-out under EU copyright law. The foundation added that it fully supports living artists using AI in their own work.
Joan Mitchell Foundation
Added Oct 17, 2026 3:50 pm ET
YouTube Official Blog
Since October 14, 2026, tracks where AI played a significant role, such as a model writing both the composition and the vocals, carry a label right on the YouTube Music player, and tapping it explains more. YouTube says it finds these songs through disclosures from music partners plus its own detection tools, and the same labels are coming to the main YouTube app.
YouTube Official Blog
Added Oct 17, 2026 3:50 pm ET
Bars Across
In a motion filed October 14, 2026 in federal court in California, DeviantArt argues that Crowell & Moring cannot represent car photographer EVOX because the firm once advised DeviantArt and parent Wix on the launch of its DreamUp image generator. Crowell has said its work for Wix ended in 2024. EVOX claims training datasets link to more than 100,000 of its photos; none of that has been decided, and the motion is set for a hearing on December 7.
Bars Across
Added Oct 17, 2026 3:50 pm ET
ABNewswire
Singapore-based Pixo announced on October 17, 2026 that its AI agent now reads a song’s sections, lyrics, timing, and singers and lays out a storyboard tied to the track. Its beat-aware editing places cuts on musical beats while avoiding breaks in the middle of a vocal phrase, and lip-sync can be assigned per singer, including duets, or switched off for story-driven videos.
ABNewswire
Added Oct 17, 2026 3:50 pm ET
Yahoo News
In an interview with Die Zeit published October 17, 2026, Axel Scheffler said he is considering suing over AI images in the style of his work, ideally together with other artists, and is talking to his publishers about support. He told the paper that asking ChatGPT for a Gruffalo produces something that is clearly his character with different backgrounds. OpenAI said it is looking into the claims and respects creators’ rights; a separate German suit over the Coconut the Little Dragon books is already under way.
Yahoo News
Added Oct 17, 2026 11:50 am ET
Deezer Newsroom
Deezer announced on October 13, 2026 that a new on/off setting, due worldwide before the end of the year, will drop fully AI-generated tracks, albums, and their artists from your search results. Music with limited AI help inside a human process is treated differently. Deezer says it now receives nearly 90,000 fully synthetic tracks a day and has kept more than 42 million of them out of recommendations since January 2025.
Deezer Newsroom
Added Oct 17, 2026 11:50 am ET
Tech.eu
Techno DJ Reinier Zonneveld launched GasLab this week with $1 million in pre-seed money from backers including Bitpanda’s Eric Demuth and WeTransfer founder Bas Beerens. The model trains only on music from artists and labels who opt in, and an attribution layer it calls the Digital Notary logs which tracks shaped each generation so contributors get a share of the revenue. Private testing invites come first, ahead of a later public launch.
Tech.eu
Added Oct 17, 2026 11:50 am ET
Artslink.co.za
Fak’ugesi African Digital & Innovation Festival runs through October 17, 2026 at Johannesburg’s Tshimologong precinct under the theme African Imaginaries. Evans Akanyijuka’s AI and Photography asks who gets to picture Africa when training data decides what looks legitimate, Renzo Filinich Orozco animates Origins Centre archives and teaches an installation of artificial neurons with South African birdsong, and Dagmar Schürrer’s Same Prompt, Different Planet looks at how the same prompt lands differently depending on who types it.
Artslink.co.za
Added Oct 17, 2026 11:50 am ET
VOZMI
PixVerse announced on October 13, 2026 that its music video feature is live: you upload a track, pick a visual style, and the tool builds the video around your audio, shown off in a 30-second 1080p demo. It sits on top of PixVerse’s multi-model platform, which also hosts its V6 video model and third-party models like Seedance 2.5 and MiniMax H3. Because the song is the input, the job becomes choosing a look rather than generating sound.
VOZMI
Added Oct 17, 2026 11:50 am ET
Mirelo
On October 14, 2026, Mirelo released SFX 2.0, a video-to-sound model that reads the finished clip, preserves spoken performance, strips unwanted baked-in music, and writes effects, foley, and ambience that stay coherent across cuts. Studio and the API take edits up to 10 minutes in one pass, with Premiere and DaVinci plugins plus partner access on fal and Runway. The pitch is sound after the cut is locked, so you do not regenerate the take just to fix audio.
Mirelo
Added Oct 17, 2026 7:51 am ET
Suno
Suno’s October 15, 2026 release notes open Shareable Workspaces on the web: a private song library you can invite collaborators into, limited to Pro and Premier. Premier also gets shareable custom models with twice the supported inputs, so a team can riff on the same trained style instead of trading files. It is a collaboration layer on top of the existing Create library, not a new generation model.
Suno
Added Oct 17, 2026 7:51 am ET
Figma
In its October 15, 2026 Build by Design launch, Figma moved code layers into open beta so teams can drop working prototypes onto the canvas from folders, GitHub, MCP, or a live site capture. Specs give Markdown design rules a home beside the artboards, Payload arrives inside Figma Make for data-backed sites, and the Figma agent plus MCP can generate layers, run Weave workflows, and ship visual assets from the tools you already use. Twenty new MCP partner integrations shipped with the update.
Figma
Added Oct 17, 2026 7:51 am ET
Jiji (PR TIMES)
NEW MEDIA / NEW TECHNOLOGY / NEW ARTIST opens October 17, 2026 at ISLAND in Harajuku after stops in Fukuoka and Seoul. Five artists work across painting, sculpture, VR, 3DCG, NFTs, algorithms, and game art, and the room includes ego Graphica, an AI agent trained on each artist’s ideas so visitors can talk through the work. The show runs through November 1 on Thursday–Sunday afternoons, with free admission and an opening reception on the first evening.
Jiji (PR TIMES)
Added Oct 17, 2026 7:51 am ET
fal
fal’s Reframe endpoint, powered by its post-trained MiniMax H3 Max build, changes a video’s aspect ratio without cropping by inventing the missing edges while keeping the source timing and audio. Clips from a quarter-second to five minutes are supported, with ratios from 16:9 to 9:21 and resolutions up to 1080p; an optional prompt steers the new surroundings. It is API-only on fal under minimax/h3-max/reframe-video, billed per generated second.
fal
Added Oct 17, 2026 7:50 am ET
AlphaSignal
Pika released a plugin built on OpenAI’s new Plugin Extensions, which opens a visual Pika workspace next to your ChatGPT conversation. You can generate and edit images, video, music, and sound effects there, trim clips and add transitions on a small timeline, and everything syncs back to your Pika library. It runs on Pika’s existing MCP server, and your first generation through the plugin adds 250 free credits.
AlphaSignal
Added Oct 16, 2026 7:55 pm ET
Envato
Envato added Krea’s new Krea 3 model to its AI image tool on October 16, 2026. The pitch is targeted editing: select the part that is not working, describe the change, and leave the approved product, character, or lighting alone. Edits can pull from six or more reference images plus up to four style references, and you can fine-tune it around your own characters or products.
Envato
Added Oct 16, 2026 7:55 pm ET
Nuva Lab
FastVideo, working with Nuva Lab and Reactor, released a preview of FastH3-Omni, an accelerated version of MiniMax H3 that takes reference images, voice or music clips, and first and last keyframes. It outputs 5 to 15 second clips with sound already synced, and can carry motion and audio into the next clip so shots cut together. Weights and training code are scheduled for October 20, 2026.
Nuva Lab
Added Oct 16, 2026 7:55 pm ET
Eigen Radar
ulearn launched Drawspire on October 15, 2026, a browser comic tool where a rough line drawing sets the composition and written directions describe the scene. You can also start from characters you drew yourself, then place panels freely, let figures break panel borders, and add bubbles and sound-effect lettering. A similarity check flags images that may look too close to existing works, though it is not a legal ruling.
Eigen Radar
Added Oct 16, 2026 7:55 pm ET
Seoul Economic Daily
At a webtoon policy meeting in Seoul this week, cartoonists and publishers asked Culture Minister Choi Hwi-young for AI training, tool vouchers, and help pairing AI firms with studios that own comic IP. One cartoonist said nearly every working artist already uses AI. Choi said individual vouchers could cost too much to justify, but pointed to a new Culture Technology Alliance launching October 15 as a place for webtoon companies to join in.
Seoul Economic Daily
Added Oct 16, 2026 7:55 pm ET
OpenMOSS (GitHub)
OpenMOSS released MOSS-Image-Base and a fine-tuned MOSS-Image-SFT on October 16, 2026, with weights under Apache-2.0. Besides normal prompts of up to 4K tokens, you can pass a JSON layout that places each object and line of rendered text in its own bounding box, which helps with posters and other designed pieces. It runs through a Diffusers pipeline or a bundled Gradio app.
OpenMOSS (GitHub)
Added Oct 16, 2026 3:51 pm ET
MetaverseWatcher
Maxon shipped its fall release on October 14, 2026. Cinema 4D gets a long-requested procedural Scatter system for filling surfaces with plants, rocks, or crowds, plus Liquid FX for foam and spray. Autograph now includes an optional MCP server so assistants like Claude or ChatGPT can build compositions and drive Cinema 4D alongside it, while ZBrush gains a UV editor and Red Giant adds an AI relighting tool for After Effects.
MetaverseWatcher
Added Oct 16, 2026 3:51 pm ET
Disney Research Studios
Disney Research Studios and ETH Zurich posted IBMM++ on October 16, 2026, an update to their Bézier-curve motion diffusion model for character animation. Animators can pin joints at any timing, shape ease-in and ease-out with tangents, and now extend clips, set orientations, and keep editing without the motion falling apart. The team also reports less foot sliding and longer sequences than the first version.
Disney Research Studios
Added Oct 16, 2026 3:51 pm ET
Noiz AI
Noiz AI and HKUST introduced Helix World, a world model that starts from one image and streams video plus matching stereo audio at 24 fps on a single H800 GPU. You move with the keyboard and the sound pans as the camera turns, and a new prompt changes the scene mid-walk. Code is on GitHub, a preview checkpoint is on Hugging Face, and a browser demo is open, though output is still 768×512.
Noiz AI
Added Oct 16, 2026 3:51 pm ET
Business Wire via The National Law Review
MVLAND, Meitu’s AI music video generator, launched Music Ads on October 15, 2026. You upload a product image, list its selling points, then pick one of four music styles and a performer from 20-plus characters or your own portrait. The tool writes the song and visuals as a vertical clip, and you can still edit the lyrics, music, and look before posting.
Business Wire via The National Law Review
Added Oct 16, 2026 3:51 pm ET
PR Newswire (Meshy)
On October 16, 2026, Meshy launched Meshy Edit on a new in-house model. Creators select an area with a lasso or brush, describe the change in plain language, and get a new model in about a minute while the rest stays put. It is open to all registered users at 20 credits per edit and works on high-poly Meshy 6 and 7.1 drafts.
PR Newswire (Meshy)
Added Oct 16, 2026 11:56 am ET
Google
On October 15, 2026, Google said TIME put SynthID on its Best Inventions of 2026 list. SynthID stamps invisible watermarks into AI-made images, audio, video, and text; Google says it has marked more than 180 billion images and videos, and Search, Gemini, Chrome, and the SynthID Detector can help people spot that mark.
Google
Added Oct 16, 2026 11:56 am ET
The Verge
The Verge reports that Epic is pushing an AI-first approach on a delayed Disney shooter, with staff told to lean on tools such as Anthropic’s Claude and to generate art through full AI workflows. CEO Tim Sweeney separately called it a huge pivot for maximum AI acceleration, aimed at building more rather than cutting headcount.
The Verge
Added Oct 16, 2026 11:56 am ET
TechCrunch
At TechCrunch Disrupt on October 15, 2026, Bluesky expanded Attie so creators can describe websites, quizzes, games, and polls in everyday language and publish them on the AT Protocol. Builds stay tied to a Bluesky identity, can use bring-your-own or open-weight models, and ride Bluesky’s audience for distribution.
TechCrunch
Added Oct 16, 2026 11:56 am ET
Somerset House
From October 16–18, 2026, Somerset House Studios stages 3 Evenings at Wapping Hydraulic Power Station — three nights of live performance and interactive installations that echo the 1966 9 Evenings experiments. The lineup includes Lawrence Lek, Keiken, Xin Liu, and others, with an online 3ETV channel extending the event.
Somerset House
Added Oct 16, 2026 11:56 am ET
GlobeNewswire (Kling AI)
Kling AI launched Kling 4.0 worldwide on October 16, 2026. The video model generates native 30-second takes, can extend an uploaded clip to two minutes, and accepts up to 15 reference assets and 10 keyframes for frame-level control of characters and scene changes, with 10-bit HDR output. It is live on the Kling website now, with API access and Adobe Firefly integration coming soon.
GlobeNewswire (Kling AI)
Added Oct 16, 2026 7:51 am ET
ElevenLabs
ElevenLabs added Instant Soundtracks to ElevenCreative on October 15, 2026. Upload a video and the model writes original music for it, spots moments that need sound effects, and places them on the timeline at the right timestamps. It can also use the clip’s existing audio as a style reference, and the API can return music and effects as separate tracks for your own mix.
ElevenLabs
Added Oct 16, 2026 7:51 am ET
SlopTV News
ByteDance’s BytePlus released OmniHuman 2.0 on October 15, 2026. The avatar model reads tone and emotion in speech or singing to drive face and body from a single picture, with optional text prompts for actions and camera moves, and supports multi-person and non-human characters. Single shots now run up to five minutes, priced at $0.07 per second through the API.
SlopTV News
Added Oct 16, 2026 7:51 am ET
DreamFort
DreamFort opened ReCast to everyone on October 15, 2026 after a closed preview. Instead of writing prompts, you pick a pre-built scene from a feed of memes, dances, and movie moments, add a photo for each role (up to four), and get the clip back with those people in it. Browsing is free; rendering uses credits, and new scenes are added weekly.
DreamFort
Added Oct 16, 2026 7:51 am ET
TheNextGenTechInsider
A developer released Kokon, an open-source Windows and macOS app that runs Qwen-Image 2.1 image editing on your own GPU. Built on the stable-diffusion.cpp engine instead of PyTorch or ComfyUI node graphs, it installs as a single package and keeps images off the cloud. The app is AGPL-3.0, while the Qwen weights carry a non-commercial research license.
TheNextGenTechInsider
Added Oct 16, 2026 7:51 am ET
Hugging Face
On October 15, 2026, Hugging Face documented FLUX.3 Image Klein in Diffusers, so the Black Forest Labs open model can do text-to-image, multi-reference edits, and bounding-box layout control from a Modular Diffusers pipeline. Distilled and base checkpoints are on the Hub, with NVFP4 and FP8 single-file loads, LoRA training scripts for generation and editing, and a demo Space for trying it live.
Hugging Face
Added Oct 15, 2026 7:57 pm ET
NVIDIA Newsroom
Adobe and NVIDIA said on October 15, 2026 that Photoshop, Premiere, and Substance 3D are being optimized for NVIDIA’s RTX Spark superchip, with up to 2× faster AI processing, editing, color, and effects. Frame.io also gains GPU-accelerated decode and TensorRT embedding so uploaded media becomes searchable sooner, and Adobe Brand Concierge is being tuned on NVIDIA Nemotron open models.
NVIDIA Newsroom
Added Oct 15, 2026 7:57 pm ET
Galerie Met
Galerie Met in Berlin opens under the debris by Spanish artist gozo on October 16, 2026, with the show running through November 7. Gozo runs early Stable Diffusion models locally, generates low-resolution, artifact-heavy stills, then animates and collages them so pixels and seams become the material, treating degraded generative memory as something that never quite resolves.
Galerie Met
Added Oct 15, 2026 7:57 pm ET
Civitai
Civitai added Krea 3 to its online generator on October 15, 2026, so creators can upload up to six reference images, describe a change in a sentence, and keep characters or products consistent across edits. The model offers Instant and Medium thinking modes at 1K or 2K, with launch-week Buzz pricing discounted through October 22.
Civitai
Added Oct 15, 2026 7:57 pm ET
Video Rebirth (bach.art)
Singapore startup Video Rebirth is pushing BACH 1.0 Pro as a production-oriented AI video engine with director-style first-and-last-frame control, native multi-shot clips, and up to 12 image, video, or audio references that keep products and characters consistent. The company says it ranks No. 2 on Artificial Analysis’ silent image-to-video leaderboard, with API output from 720p to 4K and clips up to 16 seconds.
Video Rebirth (bach.art)
Added Oct 15, 2026 7:57 pm ET
Krea
Krea released Krea 3 on October 15, 2026, an image editing model that changes only what you ask for while keeping products, labels, faces, and outfits consistent across up to six reference images. Krea says most edits finish in under 8 seconds at native 2K, priced at $0.06 per 2K image and $0.03 at 1K, on krea.ai and through partners including fal, Replicate, and OpenRouter.
Krea
Added Oct 15, 2026 3:55 pm ET
M+
M+ in Hong Kong is showing Winds of Hong Kong, a new Refik Anadol commission that uses machine intelligence to turn public wind and environmental data into a shifting “data painting” on one of the world’s largest LED screens. It is the latest in his Winds of . . . series and runs every night from October 14, 2026 to January 7, 2027.
M+
Added Oct 15, 2026 3:55 pm ET
The AOI
The UK’s Association of Illustrators surveyed 549 illustrators for its new report, AI and the Future of Illustration. Almost half said they knew a client had picked generative AI over them, 62% had been handed AI images in briefs or moodboards, and 36% had been asked to redraw AI output; 98% said AI companies should need permission before training on copyrighted illustration.
The AOI
Added Oct 15, 2026 3:55 pm ET
Computer Arts Society
The Computer Arts Society opens Oracle Machines at BCS Moorgate in London on October 21, 2026, with curator Geoff Davis and David Link introducing the show. It mixes computer-generated text, poetry, and AI art by artists including Mario Klingemann, Sasha Stiles, Ivona Tau, and Tivon Rice with material on Joseph Weizenbaum’s ELIZA, and runs until December 8.
Computer Arts Society
Added Oct 15, 2026 3:55 pm ET
Utopai Studios (Animation Magazine)
Utopai Studios and Huace Film & TV showed a first teaser for Journey to the West – The Lost Five Hundred Years, an AI-animated series, at MIPCOM in Cannes. The prequel follows the monk Jin Chanzi across ten reincarnations as he looks for companions including Sun Wukong; the studio says scripts, storyboards, and designs drive the work, and the series is planned for 2027.
Utopai Studios (Animation Magazine)
Added Oct 15, 2026 3:55 pm ET
Black Forest Labs (GlobeNewswire)
On October 15, 2026, Black Forest Labs released FLUX 3 Image Klein, a 5.6-billion-parameter open-weights image model for text-to-image, multi-reference generation, and editing. It keeps FLUX 3 features like native 4K output and bounding boxes, ships as base and distilled versions, and is free for commercial use under the FLUX Kommunity License for companies under $5M in annual revenue.
Black Forest Labs (GlobeNewswire)
Added Oct 15, 2026 11:50 am ET
aratama 璞
Japanese artist aratama 璞 released the audiovisual work KAMI Circuit 神回路 on YouTube on October 14, 2026. Every frame came from GLSL shaders and a text renderer written by Claude Opus 5.5, with no video-generation AI or editing software, and the scenes were timed to all 454 measured beats of a Suno-made track drawing on kagura and Japanese myth.
aratama 璞
Added Oct 15, 2026 11:50 am ET
arXiv (Kandinsky Lab)
Kandinsky Lab published Kandinsky S2V 1.0 on October 15, 2026: two open speech-to-video models, Lite (2B) and Pro (21B), that animate a single portrait from an audio track and text prompt at 25 fps. In the team’s side-by-side human tests, Pro matched or beat two commercial avatar services on average, most clearly on lip sync, and the code and weights are out under the MIT license.
arXiv (Kandinsky Lab)
Added Oct 15, 2026 11:50 am ET
The Economic Times
At the CII Big Picture Summit in Mumbai, JetSynthesys launched JetKreate on October 15, 2026, a platform that takes creators from script and storyboard to finished ads, music videos, trailers, and Reels. It runs on models including Google’s Veo and ByteDance’s Seedance plus Indian models, with watermarked free credits and paid plans.
The Economic Times
Added Oct 15, 2026 11:50 am ET
Picsart
Picsart added AI 3D modeling to its Flow canvas on October 12, 2026. The new 3D node builds a textured model from text, one image, or up to five views of an object in 30 seconds to 3 minutes, lets you orbit it on the canvas, and exports a GLB plus a rendered poster image for games, web viewers, or 3D printing.
Picsart
Added Oct 15, 2026 11:49 am ET
Figma
On October 15, 2026, Figma launched Specs in open beta for paid plans: a collaborative Markdown container that lives on the canvas beside designs, FigJam diagrams, and code layers. Agents can draft and refine specs, read comments, and hand the linked context to coding tools through Figma’s MCP server, so design intent stays next to the work instead of in scattered local files.
Figma
Added Oct 15, 2026 7:52 am ET
DEV Community
On October 15, 2026, Topview published Topview 3D Builder as open source: a local-first workflow where coding agents and a Python CLI place characters, props, and cameras, then render PNG reference frames. The company says the scene stays editable on your machine under Apache-2.0, so you can fix blocking before spending credits on an AI video pass.
DEV Community
Added Oct 15, 2026 7:52 am ET
Deadline
Deadline reports that VN.ai’s first quarterly copyright stress test, shared October 14, 2026, found Midjourney v8.1 the worst of 15 models at reproducing studio characters, with brand-safety violations on about 37% of prompts. Every model tested produced at least one protected character even when the name was left out of the prompt; Midjourney did not comment, and studio lawsuits against it remain open.
Deadline
Added Oct 15, 2026 7:52 am ET
Creativly
On October 9, 2026, Creativly introduced Creative Agent, a new app where you attach a sketch, product photo, or logo, talk through changes, and keep results on the same Assets and Canvas views. Direct image and video controls sit beside the agent chat, and starter cards cover mockups, packaging, and room makeovers, so a follow-up edit does not force a jump to a separate tool.
Creativly
Added Oct 15, 2026 7:52 am ET
EIN Presswire
FLASH-INTEGRO released VSDC 11.3 on October 13, 2026, for Windows, with an AI Video Generator and text-based image editing that send prompts to third-party models and drop the result on the timeline. The update also expands Garmin heart-rate and cadence fields, DJI route and altitude overlays, wide-angle correction templates, and multi-keyframe editing in the free editor’s Pro add-ons.
EIN Presswire
Added Oct 15, 2026 7:52 am ET
TechFlow
MiniMax launched Design Agent V3 on October 14, 2026, saying the update makes creation more than 30% cheaper while getting faster and smarter. Assets can now carry over between conversations, context is shared inside a single creation page, and users can write custom agent prompts so results lean toward their own style. Members can also claim 1,000 free credits a day through October 16.
TechFlow
Added Oct 14, 2026 7:55 pm ET
Art Magazine
The 15th Patchlab Digital Art Festival runs October 15–18, 2026 at Kraków’s Potocki Palace Gallery under the title Dream Sim. Highlights include Vladimir Nefedov’s Infinity Eggs, built from glitchy early AI video models, Birk Schmithüsen’s pair of sound-to-image and image-to-sound networks shown in LEDs, and fuse*’s Onirica, a dance piece where diffusion models turn fragments of more than 28,000 recorded dreams into live imagery.
Art Magazine
Added Oct 14, 2026 7:55 pm ET
V&A
Degrees of Disruption opens October 16, 2026 at V&A South Kensington as part of the museum’s Digital Art Season. The free display gathers more than 50 works, from Molnar plotter drawings like Interruptions (1969) and (Des)Ordres (1974) to pieces by Casey Reas, Licia He, Harm van den Dorpel, and William Mapan, showing how her early algorithmic experiments still shape digital art.
V&A
Added Oct 14, 2026 7:55 pm ET
Machines of Loving Grace
From October 17–25, 2026, the Karimoku Research Center in Nishiazabu, Tokyo hosts Machines of Loving Grace, an experimental show asking what it would mean for powerful technology to support human flourishing. Artists Daito Manabe and Yusuke Kamiyama used Claude to push their practices further while keeping the creative calls their own, and the exhibition adds question cards for visitors.
Machines of Loving Grace
Added Oct 14, 2026 7:55 pm ET
Visual Methodologies Collective
The Visual Methodologies Collective opens Slow & Small AI at MU Hybrid Art House on October 17, 2026, with free entry during Dutch Design Week and a run through January 3, 2027. Six artistic and design research projects, including work by Persijn Broersen and Margit Lukács, Dan Xu, and Gabriel Pereira, use local AI systems, interactive installations, experimental search tools, and speculative storytelling machines.
Visual Methodologies Collective
Added Oct 14, 2026 7:55 pm ET
Magic Patterns
On October 13, 2026, Magic Patterns released MagicOS, an infinite design canvas where PRDs, inspiration images, design systems, rejected directions, and final picks stay visible side by side. A coordinator chat spins parallel agent threads while you pan the canvas, talk by voice, and import work from Figma, Claude Design, or a codebase before handing a winner to a coding agent over MCP.
Magic Patterns
Added Oct 14, 2026 3:55 pm ET
PR Newswire
On October 14, 2026, Wizstar said its e-commerce agent builds marketing videos from existing product images, links, and other assets so catalogs need fewer dedicated shoots. The New York launch also pairs AI avatars with lip sync, viral-format recreation, and translation tools so the same clip can move across channels and languages, with human review still required for accuracy and ad compliance.
PR Newswire
Added Oct 14, 2026 3:55 pm ET
PR Newswire
On October 14, 2026, Anamana announced Anamana Japan with Tokyo’s Marketing Monster Inc., giving Japanese writers, directors, and studios access to Anamana Studio’s agentic production tools plus funding through a local Anamana 100 incubator. An early slate spans revenge, dark fantasy, and BL themes, with titles set to premiere on the Anamana app in fifteen languages.
PR Newswire
Added Oct 14, 2026 3:55 pm ET
Runway
Runway’s MCP connector lets agents generate images and videos with models such as Gen-4.5, Seedance 2.5, and Kling without leaving the chat or project. Setup is a signed-in Runway account plus the MCP URL—no API key—and generations spend the same credits as the Runway app, with Replit users adding it under Integrations for in-project web-design assets.
Runway
Added Oct 14, 2026 3:55 pm ET
Ludo.ai
On October 13, 2026, Ludo.ai added Hydra+ to its Sprite Generator: the same motion and timing as Hydra, plus a final pass that keeps engraved metal, patterned fabric, and thin line art sharp across 3–5 second clips. Improve Details applies that pass to an existing Hydra animation for one credit per second, and Hydra+ is available in Animate, Keyframe Animation, and New Direction at four credits per second.
Ludo.ai
Added Oct 14, 2026 3:55 pm ET
Linearity
On October 13, 2026, Linearity launched Linearity AI and Design Mode on the web and in Curve 7.0. One prompt returns a real design file—live text, vectors, colors, and layers—with brand rules applied from the start, then resize and translate without flattening. The company also put Curve on iPhone and opened cloud.linearity.io so campaigns are not locked to Mac and iPad.
Linearity
Added Oct 14, 2026 11:51 am ET
PR Newswire
On October 14, 2026, Pollo AI released Pollo Recast, which recreates a whole clip from a reference video’s motion, pacing, and camera work while swapping in the user’s characters, products, or scenes. The company positions it against single-layer face or style swaps, and says a viral dance, product reveal, or meme format can be remade without starting from a blank prompt. It is live at pollo.ai/recast.
PR Newswire
Added Oct 14, 2026 11:51 am ET
PR Newswire
PearlMountain Limited launched CutBay on October 14, 2026, as an AI video agent that turns prompts, images, clips, presentations, documents, spreadsheets, and links into a draft video through conversation. Users can revise script, scenes, and pacing in chat, then fine-tune text and layout on a canvas editor. Templates cover marketing, education, data viz, and business communications; the product is at cutbay.ai.
PR Newswire
Added Oct 14, 2026 11:51 am ET
Midjourney
Midjourney’s October 7, 2026 Alpha changelog, posted October 8, highlights shared folders on alpha.midjourney.com: invite Collaborators who generate into the folder or Viewers who only browse, or share by link with private, link-view, or open-to-members options. Images stay in each creator’s history, and Midjourney notes that sharing does not change whether non-stealth images can still appear on Explore.
Midjourney
Added Oct 14, 2026 11:51 am ET
Trondheim Tech
On October 14, 2026, Trondheim Kreativ AI hosts TKAI #7 at DIGS: curator Zane Cerpina (TEKS, Meta.Morf) on judging quality and originality when work is made with AI, and artist-researcher Martinus Suijkerbuijk (NTNU) on authorship and AI agents as creative partners. Talks are in English, entry is free, and a Q&A follows.
Trondheim Tech
Added Oct 14, 2026 11:51 am ET
Civitai
On October 13, 2026, Civitai said HeyGen Video is live in its generator: text-to-video, image-to-video, or reference-to-video, with dialogue, ambience, and effects generated in the same pass as the picture. Clips run 5 to 15 seconds in five aspect ratios, and Civitai is offering 50% off through October. The company calls it HeyGen’s first general-purpose video model.
Civitai
Added Oct 14, 2026 7:56 am ET
Runway Dev
Runway’s October 13 API changelog says Kling is now available on Runway Dev alongside Runway’s own models. Developers can call Kling 2.5 Turbo Pro, Kling 3.0 (Standard, Pro, and 4K), and Kling O3 through the same text-to-video, image-to-video, and video-to-video endpoints, with optional audio, multishot prompts on Kling 3.0, and up to seven reference images plus short clip edits on O3.
Runway Dev
Added Oct 14, 2026 7:56 am ET
PR Newswire
On October 9, 2026, Wondershare launched Filmora 16 under the line “Edit as an Expert with Filmora & AI.” The update tags and organizes footage with AI Media Analysis and a Smart Media Bin, adds camera tracking and color masks for local control, and lets editors reframe 360° video with keyframes and camera-movement presets. It also expands captions, subtitle tracks, and batch export for multi-version delivery.
PR Newswire
Added Oct 14, 2026 7:56 am ET
Fast Company
Fast Company reports that digital art institution Le Random has launched The Living Timeline, an interactive history with 1,147 moments from about 71,000 BCE to 2026. It opens with ochre crosshatching at Blombos Cave and runs through looms, early video processors, and today’s generative systems, arguing that generative art is any practice where a system helps complete the work—not only computer code.
Fast Company
Added Oct 14, 2026 7:56 am ET
MeiGen AI
In an October 14, 2026 launch note, Newark-based MeiGen AI described a browser workspace where one prompt bar handles image edits, text-to-image, and selected video jobs across models such as GPT Image 2, Nano Banana Pro, Seedream, Veo 3.1, Sora 2, and Seedance. Before each run it shows that model’s ratio, resolution, and credit cost, and finished files land in a My Creations library.
MeiGen AI
Added Oct 14, 2026 7:56 am ET
SlopTV News
Announced October 12, Code Video Model has a coding agent write a shot as a Three.js program, renders a rough low-poly preview in under a minute, then has the unmodified MiniMax-H3 model turn that preview plus a reference image into the final clip. Camera paths, timing and physics become lines you can edit instead of prompt guesses. It is a developer research release on GitHub (Linux and an NVIDIA GPU required), and the benchmark scores so far come from the team’s own tests.
SlopTV News
Added Oct 13, 2026 7:52 pm ET
EDAILY
An AI Media Art Zone at 2026 Content Universe Korea, October 16–17 at the Goyang Flower Exhibition Hall, will show eight new works made with generative AI under the theme “Eight Artists, Eight Perspectives.” Pieces include Kim Jin-young’s folk-painting and Dancheong patterns blooming in digital space, Park Han’s five-color fabric patchwork stitched with glowing thread, and Baek In-gyeom’s dots and lines that respond to music.
EDAILY
Added Oct 13, 2026 7:52 pm ET
Barcelona Cultura
From October 13 to 25, the Casal de Barri Font d’en Fargues in Barcelona is showing the images from the Colectivo de Autoras de Cómic campaign against generative AI, part of the Empodera’t festival. The collective says it is not against AI as a tool, but wants guarantees and authors’ rights protections over how copyrighted work is used to train models. Entry is free, weekdays 10:00 to 20:00.
Barcelona Cultura
Added Oct 13, 2026 7:52 pm ET
EIN Presswire
In an October 13 announcement, One2X’s Medeo said it is expanding from an AI video generator into one conversational workspace that suggests what is trending in a creator’s niche, generates publish-ready video, distributes it across platforms, and lets creators earn through a built-in marketplace. The claims are the company’s own; no pricing or independent review came with the release.
EIN Presswire
Added Oct 13, 2026 7:52 pm ET
OpenAI
In an OpenAI customer story published October 13, Higgsfield AI co-founder Alex Mashrabov says GPT‑6 Astra pushed the platform’s video length from about one minute to more than ten, partly by letting its creative agents call 3D tools to keep long sequences consistent. He also credits GPT‑Image‑2.5 with reliable on-image text for small-business promos, and says the next goal is “voice to movie,” where a spoken idea renders as it is described.
OpenAI
Added Oct 13, 2026 4:01 pm ET
TechRadar
TechRadar gave Gemini’s new default image model, Nano Banana 2.1, and ChatGPT Images 2.5 the same four prompts, from a greenhouse inside a snow globe to a four-panel fox comic. Google got exact poster text and a consistent character right, while ChatGPT made richer scenes and followed a detailed still-life brief more closely. The reviewer calls it near parity; 2.1 is the default on paid Google AI plans, while free users stay on Nano Banana 2.
TechRadar
Added Oct 13, 2026 4:01 pm ET
AI and Ethics (Springer Nature)
In an open-access essay published October 12 in the journal AI and Ethics, UNSW researcher Deborah Lupton presents 11 collages from her series Visualising the Language of Generative AI, pairing public-domain heritage images with phrases such as “AI slop” and “thirsty AI.” She makes them with ordinary editing tools and no generative AI, arguing that collage can puncture both the hype and the generic stock imagery that surround the technology.
AI and Ethics (Springer Nature)
Added Oct 13, 2026 4:01 pm ET
AICoder
Vivix has opened a limited API preview of Vivix-W1, which takes text, image, audio and video references and accepts new inputs mid-generation, so a scene continues and changes instead of restarting. AICoder reports the vendor’s own figures: about 30B active parameters, 5–15 second clips, and a $0.003-per-second promo price through October 31. There is no independent quality benchmark yet, and the real-time streaming API is still marked as coming soon.
AICoder
Added Oct 13, 2026 4:01 pm ET
PR Newswire
Drawing-tablet maker ugee launched ugee maker on October 12, a moderated 3D-printing community for families and beginners. Its CreateLab includes AI Magic, built with Meshy AI, which converts an uploaded sketch into a print-ready model, alongside a drag-and-drop customizing mode and a pre-tested STL library that works with standard consumer printers.
PR Newswire
Added Oct 13, 2026 4:01 pm ET
The Cool Down
Hobart crochet artist Lucy Cooper started hearing from buyers that a pattern for her spiral beanie made no sense, then found another Etsy shop selling an AI-generated version of her design with AI-made product photos. That shop listed 67 products and showed 720 sales. Etsy says sellers may use AI as a tool but not to deceive buyers or copy others, while admitting moderation at its scale is still a work in progress.
The Cool Down
Added Oct 13, 2026 12:05 pm ET
Variety
At a Kling AI session during the Busan International Film Festival market, Kim Sung-soo described going from AI skeptic to using it on Worm Train, an eight-minute short made with a nine-person team that mixed generated shots with conventional filmmaking. He plans the same hybrid approach for a seven-part drama, previewing a costly highway chase with AI before deciding what to shoot live. Panelists from MooAm, MBC, and the animated feature Born of the Tide stressed that directing skill still decides the result.
Variety
Added Oct 13, 2026 12:05 pm ET
Variety
Tom Walsh of TW3 Entertainment told Variety his two-hour holiday fantasy, based on a 2012 children’s book, was made for about $1.5 million instead of an estimated $80 million. Working with the AI studio Phatty Acid, the roughly 50-person crew had actors perform the script, an artist draw storyboards, and a composer score it, then used tools such as ElevenLabs, Seedance, and Nano Banana to generate the images and voices over a 10-month production.
Variety
Added Oct 13, 2026 12:05 pm ET
Codex Insider
An indie developer has shared Warp Studio, an animation tool with built-in MCP support, the connector standard that lets agents like Codex or Claude Code operate another program. It offers four workflows: auto-rigging and animating a 2D character, animating a single image with pins and tweening, building editable animated scenes, and deforming imported 3D models without traditional rigging. The developer says it is still in active development and lists it on itch.io.
Codex Insider
Added Oct 13, 2026 12:05 pm ET
POLITICO
A new YouGov poll of about 1,500 Australians, commissioned by the AI-attitudes research practice Diffusion, finds voters prefer copyright changes that send some money to creatives over AI companies training their models offshore for free. It lands as a bipartisan parliamentary inquiry urges Prime Minister Anthony Albanese to make it easier for AI firms to train in Australia, while rights-holders push back against loosening the rules.
POLITICO
Added Oct 13, 2026 7:52 am ET
Nintendo Life
Yuji Horii posted a free browser game, Interrogation Room Komiya, where you pick prompts to push a suspect toward a confession, and said he made it in about half an hour with AI. He compared it to building The Portopia Serial Murder Case alone decades ago, when he drew the pictures and wrote the code himself. The post drew a heavy backlash from fans and artists worried about what it means for creative jobs and for Dragon Quest XII.
Nintendo Life
Added Oct 13, 2026 7:52 am ET
Fenestra
The architecture rendering tool Fenestra has turned on Crit Space for every account. A whole project lives on one board of linked nodes: renders, variations, camera angles, video, 3D models, and team notes, with every result tied to the prompt and inputs that made it. You can drop in .3DM, .FBX, or .GLB models and set the sun, turn one image into a nine-shot set of details and angles, or animate between a start and end frame for day-to-night and seasonal changes.
Fenestra
Added Oct 13, 2026 7:52 am ET
Variety
The Painted Soul, a 10-episode animated fantasy romance from Mouse director Choi Junbae, launches exclusively on Rakuten Viki on October 21 across the Americas, Europe, the Middle East, Oceania, India, and Southeast Asia. Co-produced with Seoul studio Imaginus, whose artists used AI and other tools to build its illustrated look, it follows a trainee painter who frees a man trapped inside a painting for 50 years.
Variety
Added Oct 13, 2026 7:52 am ET
SEAGER
Natural Rhythm and Other Moral Panic, curated by Lucy Soni, runs at SEAGER in London from October 16 to November 5. Its eight artists start with analogue, physical making and then put the work through digital and computational processes, so the pieces are informed by technology rather than built by it. The show sets out to ask how artists can keep the human hand central as generative tools speed up.
SEAGER
Added Oct 13, 2026 7:52 am ET
TechCentral
Reverse Futures at the Joburg Contemporary Art Foundation in Forest Town asks whether AI and automation are being built on the same patterns of exploitation that shaped the past. Its centrepiece is Calculating Empires, a huge visual map by AI researcher Kate Crawford and artist Vladan Joler tracing technology and power since 1500, shown alongside robotics, VR, and space-themed work by artists from Africa, South America, the Middle East, and Europe. It runs until December 5.
TechCentral
Added Oct 12, 2026 7:49 pm ET
Taiwan Ministry of Culture
Taiwan’s Ministry of Culture opened the main exhibition of the 2026 Taiwan Technology x Culture Expo, Yes AI Do. Welcome to the 2050 Intimacy Exchange, at Kaohsiung’s Pier-2 Art Center. Under the theme All In AI, the expo pairs it with five other themed shows, including XR Dreamland and a MUTEK collaboration, plus 25 international forums, and asks visitors to think about relationships, love, and identity as technology speeds up. It runs through October 26.
Taiwan Ministry of Culture
Added Oct 12, 2026 7:49 pm ET
CATALOGUE
Dataland, Refik Anadol’s AI art museum at The Grand LA, launched a visitor app built with the studio Kanso, with live 3D maps and an AI guide called RAIN that answers questions as you walk through. Anadol says the museum has hosted more than 50,000 people since it opened in June; its first show, Machine Dreams: Rainforest, stays on view through January 31, 2027.
CATALOGUE
Added Oct 12, 2026 7:49 pm ET
Billboard
DistroKid confirmed to Billboard on October 12, 2026 that it removed a very small number of recordings in response to claims from Universal Music Group, which sued the distributor in September for allegedly flooding streaming services with mass-produced AI tracks and infringing remixes. DistroKid says it strongly disputes the allegations and is trying to limit disruption for affected independent artists.
Billboard
Added Oct 12, 2026 4:02 pm ET
Vixel
Vixel introduced its AI filmmaking workspace on October 12, 2026: you talk an idea through with an AI Director, which helps develop the story and break it into a storyboard before any images or video are generated. You can bring your own drawings as character references, and Vixel says its own one-minute intro grew from a 38-shot storyboard finished with human editing, voiceover, and music.
Vixel
Added Oct 12, 2026 4:02 pm ET
MANA
For the third Shanghai International Light Festival, the Zhangjiang City Lighthouse AI Art Exhibition is projecting work by 14 creative teams, picked from more than 100 applicants, onto the twin towers of the Zhangjiang Science Gate. Each piece starts from a question about cities, nature, or people and AI, shown in three weekly chapters through October 29, with a public vote deciding a Future Creator award.
MANA
Added Oct 12, 2026 4:02 pm ET
Variety
At SCAD AnimationFest, Variety’s 2026 class of 10 Animators to Watch, with artists from Cartoon Saloon, Pixar, Laika, Disney, and Sony, talked about where AI fits. Cartoon Saloon director Louise Bagnall said AI drifts toward the generic, so artists should lean into their own point of view; others pointed to collaboration and the human hand as what keeps the work worth watching.
Variety
Added Oct 12, 2026 4:02 pm ET
Vertical Slice
Vertical Slice reported on October 12, 2026 that Unity is itemising Asset Store packages for its new Spark generative platform by default, so publishers must exclude packages or items if they do not want them used there. Unity says itemisation is not training data and that publishers are emailed before content goes live; some creators still want the process to be opt-in.
Vertical Slice
Added Oct 12, 2026 12:14 pm ET
Krea
Krea’s After Effects extension docks Krea Agent beside the composition so motion designers can prompt for backgrounds, plates, motion concepts, aspect-ratio resizes, and layer cleanup without leaving the app. Results land as ordinary media assets; macOS 12+ and Windows 10+ with After Effects 2025 or newer are required, plus a Krea account and internet.
Krea
Added Oct 12, 2026 12:14 pm ET
Creative Brands Mag
At NXTinAI in Hyderabad on October 11, 2026, Galleri5 AI Studio showed 4K output, AI VFX, avatars, and agentic filmmaking pipelines aimed at long-form and theatrical work. The studio already used AI-native storytelling on Star Plus television and is developing Chiranjeevi Hanuman with Star Studio18 as a full-length AI feature.
Creative Brands Mag
Added Oct 12, 2026 12:14 pm ET
The River Reporter
How to Feel Real: Art, Technology and Us opened October 10, 2026 at MO Museum in Vilnius, co-developed with ZKM Karlsruhe. More than forty works from the 1990s to now look at AI, automation, robotics, and biotech; opening guest Edward Warchocki, a Unitree G1 humanoid with a public persona, moved through the galleries as a living prompt for the show’s questions.
The River Reporter
Added Oct 12, 2026 12:14 pm ET
Meshahara
Meshahara: Many Roots, One Ledger opened October 10, 2026 at REX KL in Kuala Lumpur with free entry by registration. Twenty-four artists from Brazil, Indonesia, Germany, Mexico, Jamaica, the USA and elsewhere show generative code, AI video, painting, sculpture, and immersive work that keeps individual authorship visible inside a shared digital ledger.
Meshahara
Added Oct 12, 2026 12:14 pm ET
Midjourney
On October 8, 2026, Midjourney opened an early Thinking Mode on alpha.midjourney.com. From a job’s lightbox, “Rerun (Thinking)” has the model re-check the image for missed objects, layout, anatomy, or text, then generate again. The company says tests show better prompt accuracy, typography, and coherence, and it is asking users where the mode helps or fails.
Midjourney
Added Oct 12, 2026 8:00 am ET
Designious
Designious put image generation into its Studio beta on October 12, 2026. Makers can start from a written prompt or a reference image, compare outputs, make variations or edits, and send a chosen result into Remove Background, Upscale, or Vectorize without rebuilding the workflow.
Designious
Added Oct 12, 2026 8:00 am ET
SlopTV News
Higgsfield announced Katana on October 8, 2026: an AI video edit tool that runs inside Anthropic’s Claude through Higgsfield’s MCP connector. Upload a reference clip, ask Claude to use Katana, and it rebuilds shot structure and pacing as editable motion graphics. It needs a Higgsfield account plus Claude Motion access.
SlopTV News
Added Oct 12, 2026 8:00 am ET
Times of Malta
Tomorrow’s Blossoms: Hybrid Thinking runs at Christine X Art Gallery in Sliema through October 24, 2026. Artist Selina Scerri and AI researcher Angelo Dalli spent five years mapping emotion hashtags into painted “semiotic alphabets,” then blending those with Stable Diffusion and other models into prints, glass, wallpaper, and a short AI-animated film.
Times of Malta
Added Oct 12, 2026 8:00 am ET
Y Combinator
YC-backed Voyager is a desktop agent for creative work: it runs on your machine and steers 100+ image and video models plus tools such as Blender, DaVinci Resolve, Ableton, and Unity. The Moda team pitches it as one place to brief storyboards, edits, and sound without juggling separate apps and API keys. macOS on Apple silicon is current; Windows and Linux are alpha.
Y Combinator
Added Oct 12, 2026 8:00 am ET
AI/CC
In a new AI/CC interview, Japan-based AI artist Toriaezoo talks about turning everyday objects into slightly unsettling, carefully balanced compositions inspired by minimal design and product photography. She argues that creative direction stays essential when working with AI, and credits a simple "let's give it a try" attitude for how her style keeps evolving.
AI/CC
Added Oct 11, 2026 7:56 pm ET
Design You Trust
Brett Keyes, who works as Color out of Place, makes vivid landscapes that start as calm lake and hill scenes and then break apart into swirling bands of color and line. He describes AI as a collaborator for exploring unfamiliar visual territory, with series such as Rippling Reality and Surreal Blues.
Design You Trust
Added Oct 11, 2026 7:56 pm ET
remio
Researchers Graelin Mandel and Alex Imas ran auctions on art and poetry made with different amounts of AI help and found that even small machine involvement lowered bids. Scarcity lifted human-made work much more than AI-made work. The paper is still an unreviewed working paper, revised on October 6, so treat the numbers as early.
remio
Added Oct 11, 2026 7:56 pm ET
Hunger
Photographer Rankin's second exhibition made entirely with AI has opened at A. Galerie in Paris, with every image picked by gallerist Arnaud Adida and no photographs on the walls. The show ranges across portraiture, fashion, mythology and art history, and asks what stays uniquely human when a machine can produce any picture. It follows the mixed reception of his 2025 London show, FAIK OFF.
Hunger
Added Oct 11, 2026 3:47 pm ET
Nasravi AI Daily
Alibaba's Qwen team released Qwen-Image-2.1-Turbo on 9 October, a 7B model that generates and edits images at 2K in eight denoising steps instead of about forty, with reference-image editing and multi-image composition. It runs on a gaming GPU, but the weights carry a research license, so studios should check terms before using it for client work.
Nasravi AI Daily
Added Oct 11, 2026 3:47 pm ET
ElevenLabs
ElevenLabs' ElevenCreative now lets teams save colors, fonts, logos, reference images, and usage rules as one Brand Kit, then tag it with @ in Image & Video, Flows, or its agent instead of re-describing a look in every prompt. Agencies can keep a separate kit per client, and voice and sound kits are planned.
ElevenLabs
Added Oct 11, 2026 11:47 am ET
Distilled News
Speaking at the Busan International Film Festival, a Netflix production director said AI has been used at various stages on roughly 300 titles, while a Korean director estimated generative AI could cut costs on his next action film by about 40%. Speakers from Netflix and Wētā FX stressed that human judgment should stay central. The same roundup notes Google's new SynthID checker for watermarked media.
Distilled News
Added Oct 11, 2026 11:47 am ET
Anime News Network
Organizers of Tokyo's Yuriten yuri art exhibition called off a planned October 17–18 kickoff event in Ikebukuro after manga creators who had shown there criticized AI-generated promotional art and decorations. Operator Lilium apologized and says future Yuriten shows and promotions will not use generative AI.
Anime News Network
Added Oct 11, 2026 7:53 am ET
Research Today
Two experiments with 600 participants, published October 9 in Frontiers in Psychology, found people felt most creative when AI contributed but they still made real choices about prompts, styles, and the final result. Fully automated generation scored lower, and moderate AI involvement gave the strongest sense of ownership and flow.
Research Today
Added Oct 10, 2026 7:54 pm ET
BBC News
Ros Atkins looks at whether audiences care if art is made by people. He cites Deezer figures that 44% of uploaded tracks are AI-generated and most listeners cannot tell, yet over half say they care, and notes "AI-free" labels starting to appear on books, films, and albums.
BBC News
Added Oct 10, 2026 7:54 pm ET
Malay Mail
The Taiwanese-born artist and curator, known as Viki Lulu, opened Art Without Boundaries: New Voyage to Nanyang at Ruang Teduh in The Exchange 106. The show pairs AI-assisted works with pieces by younger artists, runs until November 11, and aims to build two-way exchange between Chinese and Southeast Asian art scenes. Lu says AI is a collaborator, though human creativity still cannot be replaced.
Malay Mail
Added Oct 10, 2026 3:49 pm ET
CoinsTusk
Artist Dalos Dov releases 150 participatory, time-based AI artworks built from memories gathered from strangers, friends, family, and the artist. Each piece starts as a recognizable moment and keeps transforming over seven years depending on whether its owner engages with it. 105 works are offered across presale and public sale starting October 15 at 5 PM UTC.
CoinsTusk
Added Oct 10, 2026 3:49 pm ET
The Verge
Nikon pulled first place from Dr. Ning Xu's microscope video of beating cilia after scientists questioned whether it was real, ruling it broke the contest's generative-AI rules. Xu said AI was only used in post-processing; Nguyen Nam Nhat's roundworm video now holds first place, and Nikon says it will revisit its rules and judging.
The Verge
Added Oct 10, 2026 11:46 am ET
TechToday
The German singer and actress used AI in producing her multimedia project MOTHERS and is answering critics head-on. Her position: the concept, emotion, and cultural context came from human experience, and the software worked as an instrument rather than a collaborator.
TechToday
Added Oct 10, 2026 11:46 am ET
Design You Trust
The AI artist mixes anime references, chrome futurism, vampires, and Tokyo nightlife in series such as Chrome Runner and Kiss Kiss Dead Dead. Titles hint at a plot without explaining it, and side projects like a logo reel nod to Akira and classic movie-title design.
Design You Trust
Added Oct 10, 2026 11:46 am ET
Facy.ai
Facy.ai says an upcoming paid feature will animate one uploaded image into a video of up to 15 seconds at 720p or 1080p, about three times the length of typical short AI clips. Creators can write one prompt or a step-by-step sequence of what happens, with an optional prompt expander; the company says rollout timing and credit costs may still change.
Facy.ai
Added Oct 10, 2026 7:56 am ET
TechRadar
In a TechRadar interview, the #DrawWithRob creator and 2025 British Book Awards Illustrator of the Year says AI-made posters are crowding shops and cafes and taking paid work from human designers. He accepts AI is here to stay but argues companies should ask permission and pay before using creative work, rather than relying on opt-out rules.
TechRadar
Added Oct 9, 2026 7:54 pm ET
China Daily HK
China Daily profiles creators using generative tools: AI idol Yuri gave her first live show at a Shanghai AI Art Festival alongside a screening of more than 100 AI-made films, a veteran music producer pivoted to AI songs and videos after discovering Suno, and a painter uses AI for fresh ideas and to spin off merchandise and digital collectibles.
China Daily HK
Added Oct 9, 2026 7:54 pm ET
Scienmag
A study in the Journal of Big Data describes MM-DLN, a system that scores how well a drawing is composed by combining the image with written descriptions and scoring rubrics. The authors report clearly lower error and closer agreement with expert scores than earlier image-only methods, raising questions about automating art assessment.
Scienmag
Added Oct 9, 2026 3:53 pm ET
Latest with AI
Ben Vu, creator of the Battle Bears games and an early animator on Coraline, is inviting artists, musicians, writers, and makers to share work made with no generative AI on October 9, which he plans as a yearly event. The call lands the same day Guillermo del Toro’s Pan’s Labyrinth returns in a 3D conversion he chose to do without generative AI.
Latest with AI
Added Oct 9, 2026 12:13 pm ET
Stereogum
Grimes released “Accelerando,” made with longtime collaborator VADAKIN, from Psy Opera, a concept album about humanity and the risks of AI. Stereogum notes she says the album itself uses no generative AI and that about 70 people worked on its trailer, though she admitted using DeepSeek for the lyrics of one song about DeepSeek.
Stereogum
Added Oct 9, 2026 12:13 pm ET
Europa Circuit
Multimedia artist Sputniko! is debuting A Sky of Infinite Wishes, an immersive machine-learning installation about what our hopes for AI reveal about us, at Ars Electronica before it travels to Deep Spaces venues in 2027. She frames it as a look at how quickly people hand authority to AI systems, citing her young daughter asking assistants how the world works.
Europa Circuit
Added Oct 9, 2026 12:13 pm ET
Tablet Magazine
Tablet profiles Shira Barzilay, known as Koketit, whose flowing line drawings started as digital art placed over real photos with AI help. One of those pool fantasies became real when a Bali hotel commissioned her figures as a tiled mosaic, and she talks about treating every piece as a starting point rather than a finished work.
Tablet Magazine
Added Oct 9, 2026 12:13 pm ET
The Verge
Anthropic’s Claude can now add motion to what it makes, The Verge reports: ask it for a 30-second explainer video, an animated chart, or a product walkthrough, and the result plays like a short clip you can save as an MP4. Anthropic is also rolling out live dashboards tied to company data and taking its Docs, Slides, and Design features out of beta.
The Verge
Added Oct 9, 2026 7:56 am ET
Westword
Artists-in-residence at Denver’s 40 West Arts District curated “Digital Monsters: Resisting AI with Real Art” at The Three Leaches Gallery, with collage, painting, video projection, and poetry about being human in the age of AI. Westword says it is one of several local events reacting to AI-generated content, and that many artists there are turning to sculpture, ceramics, and other hands-on mediums.
Westword
Added Oct 9, 2026 7:56 am ET
OpenAI
In an OpenAI customer story, Pollo AI describes Pollo Agent, a guided workflow that turns an idea into storylines, scenes, images, and short video ads using OpenAI models including GPT-Image-2.5. The company says about 30% of creator sessions start from its templates, and its Photo to Video Ads tool turns one product photo and a prompt into a social clip. This is a vendor case study, not independent reporting.
OpenAI
Added Oct 9, 2026 7:56 am ET
Unite.AI
Magnific, the company formerly known as Freepik, launched Magnific One, an image model it says settles the look of a picture, the way an art director would, before generating it from a short prompt. A companion Brand Kit feature pulls a team’s logo, colours, type, and tone from a website or a few images and applies them to new work. It is on all paid plans on the web, desktop, and mobile apps, and the desktop app offers unlimited 2K and 4K generations through October 15.
Unite.AI
Added Oct 8, 2026 1:05 pm ET
Deadline
Abundantia Entertainment founder Vikram Malhotra told Deadline that AiOn, the AI production arm the company started a year ago, will become its own native AI studio in 2027, making fully generated and hybrid films. He said he is getting about three times as many pitches as before from people outside the film business, and he calls the new studio human-first. Its slate includes Chiranjeevi Hanuman and the hybrid Jai Santoshi Mata, plus films made with InVideo.
Deadline
Added Oct 8, 2026 1:05 pm ET
Business Today
The Indian Institute of Creative Technologies and YouTube showed four AI-assisted short films in Mumbai, made by Eleventh Door Studio, Avium Studios, Karthik Kult Kreations, and Kalpanik Films. The studios got mentoring and facilities from IICT plus Google and YouTube AI tools and credits, and used them from concept through post-production. The films and behind-the-scenes videos are now on the filmmakers’ YouTube channels. The article is a sponsored feature.
Business Today
Added Oct 8, 2026 1:05 pm ET
Music Business Worldwide
The US Copyright Office published a notice on October 7 asking how music streaming fraud works, what it costs artists, and how AI-generated songs feed into it. The notice came a day after Michael Smith was sentenced to 18 months in prison for using bot accounts to stream hundreds of thousands of AI-made tracks. The office is also asking whether new laws allowing no-fault court orders could help stop the practice.
Music Business Worldwide
Added Oct 8, 2026 1:05 pm ET
The Copenhagen Post
Danish Culture Minister Zenia Stampe introduced a bill that would make it illegal to share realistic AI imitations of a person’s appearance or voice without their permission. It covers ordinary people and artists, whose performances would also be protected, and pushes social platforms to take illegal deepfakes down quickly. Satire and parody stay legal, with courts deciding where the line falls. Parliament still has to approve it, and Stampe hopes it takes effect around the turn of the year.
The Copenhagen Post
Added Oct 8, 2026 7:52 am ET
Dublin Inquirer
Illustrator Lauren Fernandes noticed her favourite ice-cream shop and café in Balbriggan, near Dublin, using AI-made posters, so she offered to draw posters for small local businesses in exchange for things like a family meal. The offer got warm replies on Facebook, but no one has taken her up on it yet. The Dublin Inquirer piece also talks to Thomas House, a Dublin bar and venue that has stopped accepting AI gig posters from bands.
Dublin Inquirer
Added Oct 8, 2026 7:52 am ET
Synthesia
Synthesia introduced Syren, a video engine that studies a company’s existing videos to pick up its colours, fonts, animation style, and pacing. You describe the video you want in a chat, and Syren builds it with motion graphics, 3D, voiceover, music, and avatars, then lets you refine it the same way. Synthesia ran a waitlist ahead of the release, with bonus credits for the first 1,000 eligible sign-ups.
Synthesia
Added Oct 8, 2026 7:52 am ET
Mint
Google launched Playground with Unity, a place to describe a game idea in plain text, build it with Google’s AI models, and publish it for others to play; Unity’s Spark tools are due to join later this year. Unity’s CEO said the aim is to sit alongside existing tools, not replace developers. Google also teamed with Women in Film India on a Filmathon where filmmakers make original films in its Flow tool, with five to be shown at IFFI Goa 2026.
Mint
Added Oct 8, 2026 7:52 am ET
80 Level
Unity announced Spark, a browser-based tool for describing, building, and publishing games without learning code or the full Unity Editor. Instead of leaning on generated visuals, it pulls from the Unity Asset Store, and the Financial Times reports the artists whose assets are used will be credited and paid. Finished games can be shared through Google’s new Playground platform. A closed beta is planned for later in 2026, and pricing is not set yet.
80 Level
Added Oct 7, 2026 7:48 pm ET
Newsfile
HiDream.ai made vivago R1 available worldwide. It is a chat-driven video tool where a lead agent plans the project and helper agents handle the script, storyboard, characters, sets, and audio, keeping them consistent across shots and even across episodes. It can make videos up to five minutes long, and a command-line tool lets other software or AI agents start jobs. The company plans to add its HiDream-O1-Video model, with built-in sound, in a later update.
Newsfile
Added Oct 7, 2026 7:48 pm ET
SlopTV News
Spira AI released Spira Maxima 1.0, which takes a script, and optionally your photo, voice, or product shots, and returns a finished 9:16 social video of up to 90 seconds with a presenter, B-roll, captions, and music. It costs $0.15 per second of video and has a free tier. Spira’s agents can also post to TikTok, Instagram, and X by themselves, and Product Hunt users asked how the face and voice cloning handles consent. The reach numbers it promotes are the company’s own.
SlopTV News
Added Oct 7, 2026 7:48 pm ET
PetaPixel
Software engineer Brandon Thomas has put out ArtCraft, an alpha suite of free, open-source stand-ins for seven Adobe apps, including PhotoCraft, FilmCraft, and LightCraft, for macOS, Windows, and Linux. He says he used an AI coding model to rebuild them in Rust as a clean-room reimplementation and is aiming for feature parity within a month. The apps work without a subscription, with optional paid credits for built-in generative tools, and patent questions remain open.
PetaPixel
Added Oct 7, 2026 3:55 pm ET
Unite.AI
ShengShu Technology opened a public preview of Vidu Q4, its next flagship video model, on the Vidu web app and API, with launch pricing from $0.014 per second. A single setup can take up to 15 reference images and three reference audio clips to keep characters, props, and voices consistent, and it outputs clips of up to 16 seconds, with sound, at up to 4K and 10-bit color. The company says feedback on the preview will shape the full release.
Unite.AI
Added Oct 7, 2026 3:55 pm ET
TechTimes UK
The Roblox studios behind Steal a Brainrot are suing French agency Mementum in federal court in California after it demanded they stop using the viral character. Court papers say it was made in about 15 minutes from seven prompts, and the studios argue that is too little human authorship for copyright. Mementum has answered with trademark and unfair-competition claims, and no court has ruled yet on who, if anyone, owns the character.
TechTimes UK
Added Oct 7, 2026 3:55 pm ET
Agenda Culturel
A show near Toulouse, running through October 17, looks at image-making with AI in three parts: pictures, faces, and artworks generated by AI, a head-to-head where a researcher’s AI and a human artist answer the same brief, and a collaboration where one sketches and the other finishes. It is put on by the science center Quai des Savoirs and Umeshu Lovers with the AI institute ANITI.
Agenda Culturel
Added Oct 7, 2026 3:55 pm ET
TechCrunch
Google has made its SynthID detector public at synthid.com after a year of limited testing with journalists and researchers. Anyone signed in can upload an image, video, or audio file to see whether it carries the invisible SynthID watermark that Google's own generators, including Nano Banana, Veo, and Lyria, embed in their output. It only recognizes that watermark, so art made with tools that use other systems will not be flagged.
TechCrunch
Added Oct 7, 2026 11:56 am ET
Variety
Fable, the AI film studio, pushed the U.S. theatrical release of Gods Don’t Give Gifts, four satirical shorts by YouTuber Gossip Goblin (Zack London), from October 30 to December 4 so it can run an Academy Awards campaign. The plan includes screenings for voters and critics and an LA billboard push, and it will test how the Academy treats a film made with AI, since its rules say judges weigh how much a human was at the heart of the authorship.
Variety
Added Oct 7, 2026 11:56 am ET
inITALY
MAM, a 1,000-square-meter museum devoted to media art, opens on Corso d’Italia in Rome on October 9. Its permanent show, HyperMedia, traces more than two decades of digital work from early net art through generative and AI-driven pieces and interactive installations, and a 300-square-meter room wraps walls, floor, and ceiling in projection. The opening season also includes artist residencies, the Live Performers Meeting, and a live cinema festival.
inITALY
Added Oct 7, 2026 11:56 am ET
Aalto University
Aalto University's art and design school will show art-ai-fact at Dutch Design Week in Eindhoven, October 17 to 25. The projects use AI in physical making: one generates textile patterns and textures and turns them into instructions a digital textile machine can produce, and another, LUCID, lets designers steer and edit AI-generated shapes on their way to becoming 3D-printed objects, with the system keeping a visible record of how each form evolved.
Aalto University
Added Oct 7, 2026 11:56 am ET
EIN Presswire
Printful is rolling out AI Mockups, which let merchants pick a preset model and background or describe a scene and generate lifestyle product photos during the publishing step instead of hiring a shoot or juggling outside tools. It launches for more than 95 apparel products, with more categories planned. For artists selling designs on shirts and hoodies, it means showing the work on a person without leaving the store builder.
EIN Presswire
Added Oct 7, 2026 11:56 am ET
MarTech Series
Gamma, the AI tool for presentations, documents, and websites, released Gamma 5, which the company calls a ground-up rebuild aimed at the samey layouts and generic imagery that make AI-made visuals easy to spot. The new version pairs a conversational agent that keeps context while you edit with a freeform canvas for moving individual elements, a much wider range of visual styles, brand settings that carry across projects, and import and editing of existing PowerPoint files and PDFs.
MarTech Series
Added Oct 7, 2026 7:54 am ET
EIN Presswire
GenWave announced SVGMaker 2.0, whose new design agent sets up a brand's colors, type, and visual style and then applies them across logos, icons, illustrations, labels, merch, and Cricut files. Results stay as editable SVG paths instead of flat images, raster art can be converted into vectors, and a browser editor offers Bézier control, layers, and boolean operations so you can tweak a shape without regenerating the whole design. It also offers an API and MCP hookup for automated workflows.
EIN Presswire
Added Oct 7, 2026 7:54 am ET
Gajah Gallery
Balinese artist Jemana Murti's second solo show, Source Code, is on at Gajah Gallery in Singapore through October 25. He uses 3D scanning, 3D printing, and AI to rebuild traditional Balinese forms with deliberate distortions that point to what technology cannot capture about their cultural and spiritual meaning. A new series made with Yogya Art Lab turns those warped digital reconstructions into bronze sculptures.
Gajah Gallery
Added Oct 7, 2026 7:54 am ET
Illustro
Illustro is a new tool for product and marketing illustration: you describe a scene, choose a style such as flat, line art, 3D, or children's book, then edit the result in the same workspace. Illustration packs lock a character first so the same person appears across a whole onboarding flow or site, and finished files can go straight onto a Figma canvas, a website, or a slide. Its makers pitch it as a replacement for stock art that is too generic to ship.
Illustro
Added Oct 7, 2026 7:54 am ET
Google
Athens-based artist Theo Triantafyllidis, this year's Frieze London Artist Award winner, is showing Feral Metaverse (Spider), a soft, giant spider sculpture that visitors help each other climb to join a live multiplayer game. Google says his team of sculptors, fabricators, and game developers used Gemini to generate sketches, character designs, backdrops, and 3D renders that shaped the game's look. It is on view at Frieze London in The Regent's Park, October 14 to 18.
Google
Added Oct 6, 2026 7:48 pm ET
Slaughter and May
The European Commission launched a targeted consultation on its copyright rules that asks how creators who opted out of AI training can actually enforce that choice, license their work, and get paid. Ideas on the table include a standard opt-out protocol with a central registry, disclosure rules for web crawlers, audits of AI providers, and compensation rights, plus a section on AI imitations of performers. The Commission says the options are not its chosen plan; responses are open until November 3.
Slaughter and May
Added Oct 6, 2026 7:48 pm ET
Dexerto
Fans noticed a Google Gemini logo in the corner of the cover art for "Meow" by Ironmouse, Sleeping Forest, and HalaCG in Fortnite Festival. Ironmouse said she had not known and her team demanded a change; by October 5 a version without the mark had replaced it. Epic Games told Dexerto the artwork came from the artist and matched other streaming platforms, and that it updated the image once the artist did.
Dexerto
Added Oct 6, 2026 7:48 pm ET
ArtRealmAI
Researchers at Hong Kong Polytechnic University and OPPO released open weights and code for PVD, which splits image generation into a quick drafting expert and a refining expert running on a slimmed backbone. Its distilled FLUX.1-dev version peaks around 12 GiB of memory instead of about 23, putting it within reach of mid-range home GPUs, with versions for Qwen-Image and SD3.5 Medium too. There is no ComfyUI node yet, and the FLUX version still carries FLUX.1-dev's commercial-use limits.
ArtRealmAI
Added Oct 6, 2026 7:48 pm ET
Google AI for Developers
Google made Nano Banana 2.1 generally available in the Gemini API on October 6. It updates Nano Banana 2 for image generation and chat-style editing, with sharper results, closer prompt following, steadier characters across edits, and cleaner text in images. It also adds very wide and tall frames, up to 8:1 and 1:8, at 1K, 2K, and 4K. The older Nano Banana 2 model is set to shut down October 29.
Google AI for Developers
Added Oct 6, 2026 3:54 pm ET
ArtRealmAI
Prism, from Fudan University and Tencent Hunyuan with Zhejiang University, generates picture and soundtrack together at 720p, 1080p, or 2K instead of upscaling a small clip. Code and two preview checkpoints are out under the MIT license, and prompts take separate parts for the image and the audio. It is built for big hardware: 720p needs one 80 GB GPU, and 2K needs four or more.
ArtRealmAI
Added Oct 6, 2026 3:54 pm ET
Hugging Face
inclusionAI posted Ming-Image-0.1-Design on Hugging Face, a 6-billion-parameter text-to-image model aimed at text-heavy design work like posters, infographics, and app mockups. It can output images with transparent backgrounds, which makes it easier to make a subject and a backdrop separately and layer them later. There is a free demo Space, and the recommended setup is 2048 by 2048 on a single 80 GB GPU.
Hugging Face
Added Oct 6, 2026 3:54 pm ET
Niagara Frontier Publications
The Castellani Art Museum at Niagara University is showing STANZA, a solo show by poet, artist, and AI researcher Sasha Stiles, after her A LIVING POEM ran at MoMA. The installation asks visitors to walk into poetry that can be heard and seen, not just read, continuing work with language models she began in 2018. The opening reception is Thursday, October 8, and the show runs through April 25, 2027.
Niagara Frontier Publications
Added Oct 6, 2026 3:54 pm ET
Trademark Lawyer Magazine
A federal judge in California refused on October 2 to dismiss a case from Ted Entertainment and creator Matt Fisher, who say ByteDance got around YouTube’s technical protections to collect their videos for AI products including MagicVideo and Seedance. The ruling only says the claim was pleaded well enough to continue; it does not find that ByteDance did it. It hints that anti-circumvention law could matter in future fights over scraping training data.
Trademark Lawyer Magazine
Added Oct 6, 2026 3:54 pm ET
MoMA Magazine
In a new MoMA Magazine conversation with curator Stuart Comer, Huyghe talks through UUmwelt, now in the museum’s sculpture garden. A person imagined objects inside an fMRI scanner at a Kyoto lab, and a neural network turned those scans into thousands of guesses at what they pictured. Visitors’ movements feed new images, and simulated cancer-cell mutations set their timing. It is on view through November 29.
MoMA Magazine
Added Oct 6, 2026 11:59 am ET
CATALOGUE
The free London premiere of Man’s performance-lecture is Saturday, October 10, at V&A South Kensington, as part of the museum’s Digital Art Season. StarQuest strings 111 short scenes made only with text-to-video and text-to-music models, drawn from American competitive dance and Dance Moms, into a stream that never plays the same way twice. Her net artwork Pure Information also recently sold for $20,000.
CATALOGUE
Added Oct 6, 2026 11:59 am ET
GosuGamers
Skin designer Kitlosi said he left the team because the game kept using AI, pointing to assets in its latest update. Developer Roytt admitted an AI filter was used to polish screenshots shown on an in-game TV, removed them, and promised to delay unfinished art rather than fill gaps with AI. Players note it is the studio’s second apology of this kind.
GosuGamers
Added Oct 6, 2026 11:59 am ET
CREAO
CREAO added a page just for making AI video from a prompt or a ready-made example, such as product shots, logo reveals, interiors, or anime scenes. You can run the same prompt through Veo 3.1, Seedance 2.0 and 2.5, or Gemini Omni Flash and keep the best take, with vertical to cinematic formats and sound. Editing existing clips with Gemini Omni Flash works on the free plan.
CREAO
Added Oct 6, 2026 11:59 am ET
eroq
eroq added Make, an agent mode that takes a goal like a short ad, a noir short, or a music video teaser and runs the production with you. It asks what it needs, writes a storyboard, and shows the cost of each render, then renders only after you approve and suggests the next step until you have a cut to open in its editor.
eroq
Added Oct 6, 2026 11:59 am ET
ArtRealmAI
Sber’s Kandinsky Lab released Kandinsky 6.0 Video under the MIT license, with code on GitHub and weights on Hugging Face. It turns a prompt or a still image into a five-second clip whose soundtrack, including lip-synced speech, effects, or music, is generated together with the picture, and an optional upscaling pass reaches Full HD. You can run it on your own NVIDIA card, try a free online demo, or add it to ComfyUI.
ArtRealmAI
Added Oct 6, 2026 7:50 am ET
Courthouse News Service
Austin Beaulier sued Roblox after the NoAI tags on his 3D models were stripped and the models were used to train its AI. A federal judge in San Jose agreed the tag counts as protected copyright management information and that he had standing, but found he had not shown the removal was intentional. He has 30 days to file an amended complaint.
Courthouse News Service
Added Oct 6, 2026 7:50 am ET
Scienmag
Researchers at Hangzhou Normal University and Shenzhen Polytechnic University showed people identical work under different labels. Artworks were rated more creative when credited to a person, yet answers to a classic brainstorming test, listing unusual uses for everyday objects, scored higher when labeled as AI. The paper appears in BMC Psychology and suggests the bias depends on what kind of creativity is being judged.
Scienmag
Added Oct 6, 2026 7:50 am ET
tiat (Luma)
San Francisco art-and-tech space tiat opens a three-artist show on October 9. With support from Google, textile artist Hayley Mortin, media artist Karyn Nakamura, and performance artist Zelia ZZ Tan dug through an archive of odd, unstable videos made while early Veo models were being tested, and made new work from those leftovers.
tiat (Luma)
Added Oct 6, 2026 7:50 am ET
100 collectors
Gallery Renaissance shows Rayner’s generative work from October 13 to 18, with an opening on the 13th. He has spent more than thirty years writing his own algorithms rather than prompting a model, aiming for the color and texture of abstract painting, and this show moves that work off the screen into large-format pieces and framed prints.
100 collectors
Added Oct 6, 2026 7:50 am ET
Nieman Lab
Nieman Lab found OpenAI’s image tool not only mimicking the magazine’s cartoon style but adding the signatures of more than 15 real New Yorker cartoonists, from Brendan Loper to Saul Steinberg, without their permission. After the reporter flagged it, ChatGPT began warning about similarity to third-party content, yet some generic cartoons still came out signed. The New Yorker says it never allowed its cartoons to be used for training.
Nieman Lab
Added Oct 5, 2026 7:54 pm ET
EIN Presswire
The Melbourne music platform updated its browser visual tool so a text prompt becomes a small animation program instead of a finished video. It listens to your track as it plays, so motion follows the beat and bass, and you can still tweak color, size, text, and how hard it reacts. It is aimed at looping vertical clips for Spotify Canvas, Reels, TikTok, and Shorts, and runs on the site’s in-app credits.
EIN Presswire
Added Oct 5, 2026 7:54 pm ET
ZKM
Listed by Germany’s ZKM, this long-running exhibition traces how artists have responded to technology, from the first computers and the early internet to today’s AI. Its themes are connection, automation, intelligence, the body, and agency, framed around what it means to feel alive in a digitally mediated world. It runs until August 29, 2027.
ZKM
Added Oct 5, 2026 7:54 pm ET
Thaddaeus Ropac
After showing them at Palazzo Cini during this year’s Venice Biennale, Salle reunites the series at Thaddaeus Ropac’s Pantin space from October 21 to December 23. He trained a personal model on his own late-’80s Tapestry Paintings and used it to break apart and reshuffle compositions, then painted the results by hand on canvas. The gallery frames it as using AI to get outside his own habits, not to replace the painter.
Thaddaeus Ropac
Added Oct 5, 2026 7:54 pm ET
inITALY
La Zaffera Art Gallery is staging a two-week exhibition built around Canto XXXIII of Dante’s Paradiso, combining AI-made short films with photographs. It is endorsed by the Italian Dante Society, opens with a reading on October 10, closes with a critical talk on October 24, and admission is free.
inITALY
Added Oct 5, 2026 7:54 pm ET
Kotaku
On the October 5 episode of his KOJI10 podcast, the Death Stranding director said he has mixed feelings about Nvidia’s DLSS 5 and AI remastering in general. His worry is intent: he chose specific lighting for a reason, and an AI that does not understand those choices could quietly rewrite old films, albums, and games. He said it might feel different if he were the one steering the AI on his own work.
Kotaku
Added Oct 5, 2026 3:56 pm ET
Reka
Reka released a research preview of Rho-1, a 19B model trained from scratch that both understands and generates text, images, and video without handing work off to separate tools. In its demo it draws a scene, boxes an object, animates it, then restyles the clip into a snowstorm. A distilled version returns short clips in about a second, though Reka notes native video is capped at 672×384 and edits are still brittle.
Reka
Added Oct 5, 2026 3:56 pm ET
Museum de Fundatie
Museum de Fundatie in the Netherlands is hosting the digital art pioneer’s largest European solo show so far, organized with Kunsthalle München. It spans early computer pieces, works made with AI, a robot that draws on site, and big interactive installations whose algorithm-made images shift as visitors move. The show treats landscape and nature as its theme and runs until April 4, 2027.
Museum de Fundatie
Added Oct 5, 2026 3:56 pm ET
TechCrunch
Starting later in October in the US, ChatGPT will place clearly labeled visual ads from a first group of test advertisers alongside images people ask it to generate. OpenAI says the ads will not influence its answers, and it is adding measurement partners plus brand-safety pilots with DoubleVerify and IAS as it builds out an ad business for free and low-cost users.
TechCrunch
Added Oct 5, 2026 12:01 pm ET
EDM.com
After fans criticized the neon dog cover for Kevin D’Angello’s October 1 single, the dance label posted an apology using the phrase “AI slop” itself, said the image skipped its normal approvals and was not made by its usual 3D design team, and swapped it out. Revealed says it is setting up a new review process and will share more of its design work in progress.
EDM.com
Added Oct 5, 2026 12:00 pm ET
Ground Truth
In an October 4 statement, Jagex said it removed a Dragon Slayer II trailer made by an outside partner that had promised not to use generative AI. The studio repeated that it does not use genAI for its games, artwork, or creative assets and will not knowingly let others use it on its behalf, and said it will tighten reviews and commission community creators instead.
Ground Truth
Added Oct 5, 2026 12:00 pm ET
Aargauer Kunsthaus
Organized with Basel’s House of Electronic Arts, the Swiss museum’s show looks at making art with AI as a push and pull between control and losing it. Artists include Botto, Terence Broad, Holly Herndon and Mat Dryhurst, Simone C Niquille, and Hochschuh & Donovan, whose 50 small robots roam together. It runs from September 26, 2026 to October 1, 2027.
Aargauer Kunsthaus
Added Oct 5, 2026 12:00 pm ET
PetaPixel
Speaking to artists and curators at the Vatican on October 2, 2026, Pope Leo XIV argued there is a basic difference, before any question of looks, between human art and images a machine computes from millions of other people’s work. He said the Church wants a renewed alliance with artists and cultural institutions, his latest public pushback on AI-made art.
PetaPixel
Added Oct 5, 2026 7:53 am ET
Music Business Worldwide
By the end of September 2026, Sony Music had asked streaming and social platforms to pull more than 260,000 AI-generated songs imitating the voices of artists such as Harry Styles, Adele, Queen, and Michael Jackson, nearly double its March tally of 135,000, according to Financial Times figures reported October 5. Sony backs the US NO FAKES Act as a fix.
Music Business Worldwide
Added Oct 5, 2026 7:53 am ET
Taiwan News
Japanese media artist Manabe Daito’s first major solo show abroad, Relational Time – Life Beyond Simulation, fills three galleries in Taichung with new site-specific works that react to visitors, body signals, environmental data, and AI agents. He says AI-assisted coding, more than 21 million lines of it, let him build ideas he could not before. On view through January 3, 2027.
Taiwan News
Added Oct 5, 2026 7:53 am ET
Tekedia
The Paris museum is forming a Digital Art Committee and a five-year patron program to fund acquisitions and research, with a focus on generative and onchain art. The commentary piece notes the hard part for museums: preserving not just a file, but the code, parameters, and software environment a generative work needs to run.
Tekedia
Added Oct 5, 2026 7:53 am ET
Trebuchet
Borusan Contemporary’s Haunted, running through August 15, 2027 at Perili Köşk, pairs works by Rafael Lozano-Hemmer, Memo Akten, teamLab, Quayola, DRIFT, and others to ask how algorithms and data now shape the images we see. In an interview, curator Kumru Eren says the point is not real versus fake, but noticing the unseen systems behind every image.
Trebuchet
Added Oct 5, 2026 7:53 am ET
Finestre sull’Arte
Running October 3 to December 8, 2026 across four venues in Matera, Italy, this exhibition curated by Chiara Canali treats the landlocked city as part of the Mediterranean. Baris Gencel uses AI to turn faces into fossils and coral, Silvio Giordano reworks Carlo Levi’s Lucania ’61 with AI, and others use computer vision, holography, and fractals.
Finestre sull’Arte
Added Oct 4, 2026 7:54 pm ET
Gray Area
San Francisco’s Gray Area and cloud company Modal picked 11 projects from a global call for AI-enabled artworks, giving each a stipend, a materials budget, and $20,000 in compute credits. The finished pieces go on view in Scaling Intelligence at Gray Area from October 28 to November 1, 2026, pitched as a counter to the idea that AI output is just slop.
Gray Area
Added Oct 4, 2026 7:54 pm ET
ZKM
Artists Daniel Heiss and Marc Schütze are bringing their ZKM installation to EnBW’s Karlsruhe headquarters from October 23, 2026 to February 7, 2027, with a free opening on October 22. Autonomous AI agents gather public data, form hypotheses about hidden links, and keep redrawing them as infographics, networks, and maps.
ZKM
Added Oct 4, 2026 7:54 pm ET
Mickey Visit
Independent theme-park apparel brand The Lost Bros posted side-by-side comparisons showing AI-generated copies of two of its most-copied shirt designs, which it says turn up for sale on Etsy and are worn around EPCOT. Mickey Visit’s October 4, 2026 write-up frames it as a shift from copying Disney itself to copying the small designers who riff on it.
Mickey Visit
Added Oct 4, 2026 7:54 pm ET
DevPik
Coverage dated October 4, 2026 describes Hypit, a free open-source agent skill that rebuilds a reference video as an editable composition—footage, captions, B-roll, and effects anchored to spoken words instead of timestamps—then ships many variants from one command. Generation models are optional; workflows can compile captions and code-rendered visuals locally. Install is via npx skills add, with documented example clips costing about a dollar each in model fees.
DevPik
Added Oct 4, 2026 3:51 pm ET
Game Up News
On October 4, 2026, Game Up News reported that Layer launched Playables, an agentic tool that turns a game title, brief, screenshots, or gameplay capture into MRAID-compliant HTML5 playable ads for networks such as AppLovin, Meta, TikTok, and Unity Ads. A first draft arrives in minutes and campaign-ready files in hours, with mechanics and art adjustable by chat. It ships in existing Layer plans with no per-export fee.
Game Up News
Added Oct 4, 2026 3:51 pm ET
ForPressRelease
On October 4, 2026, Nereo (operated by Shoshone AI Inc) announced a platform that turns text prompts, product photos, or scripts into short promotional videos, product demos, and social ads in one place. The pitch targets marketers, online stores, agencies, and small teams that need a steady stream of brand-ready clips without a full production pipeline. The company points creators to nereo.io to try it.
ForPressRelease
Added Oct 4, 2026 3:51 pm ET
arXiv
An October 2026 arXiv paper from Cornell, UC Berkeley, Harvard, and Impossible Inc. introduces MosaiChunk, a spatio-temporal memory that composes selected historical key–value sections so autoregressive video generators keep fine details when objects reappear after leaving the context window. The team freezes the generator, trains only a lightweight router, and reports stronger revisit consistency on a new RememBench of long-horizon T2V and I2V revisits. Code and data are promised for release.
arXiv
Added Oct 4, 2026 3:51 pm ET
arXiv
Johns Hopkins researchers posted Generative Cinematographer (GenCine) on arXiv in October 2026: it lifts a single image into an editable 3D scaffold where artists set camera paths and local 3D motion handles, then converts those controls into guidance maps for a pretrained Wan video model. Multiple handles can move different parts of a subject independently as a piecewise-rigid stand-in for non-rigid motion. Experiments report more consistent camera-relative motion and geometric control than 2D-trajectory baselines.
arXiv
Added Oct 4, 2026 3:51 pm ET
Creative Bloq
Creative Bloq’s October 4 review covers Adobe Premiere Mobile on Android: multi-track editing, Enhance Audio, and watermark-free 4K exports without a Creative Cloud plan. Firefly features such as Generative Fill, Image-to-Video, and Generate Sound Effects still spend generative credits.
Creative Bloq
Added Oct 4, 2026 11:48 am ET
Atlabs
Atlabs published a full walkthrough of “Makhan Chor,” a 4:59, 108-scene 3D-cartoon music film built from an uploaded Hindi track and one written concept. Storyboard mode timed each shot to lyric phrases and kept ten characters consistent across the export.
Atlabs
Added Oct 4, 2026 11:48 am ET
Atlabs
Atlabs’s October 1 guide shows how its Model Context Protocol server lets Claude pick models, show credit cost, and chain image, video, lip-sync, motion control, and music tools in one chat. Outputs keep IDs so a still can move into animation, reframe, and soundtrack without re-uploads.
Atlabs
Added Oct 4, 2026 11:48 am ET
Uisato Studio
Uisato Studio’s Oscilloscope Diffusion takes an uploaded clip—live action, animation, or audio-reactive geometry—and runs frame-by-frame diffusion so materials and style change while source motion guides the piece. Renders can take hours; output tops out at 1,024 px on the long edge.
Uisato Studio
Added Oct 4, 2026 11:48 am ET
arXiv
Stability AI researchers introduce 4Director, a video world model that rebuilds each marked object as a complete mesh, moves it with one rigid transform per frame, and renders depth so a Motion Adapter can fill in appearance and non-rigid motion. They also release RealCOD-Rigid, 20,774 annotated clips for training.
arXiv
Added Oct 4, 2026 11:48 am ET
Qazinform
On October 3–4, 2026, Kazakhstan’s first Astana AI Film Festival closed with 8,067 entries from 125 countries. Latvian filmmaker Nikolay Shestak took the Grand Prix and $450,000 for The First Honest Day; India’s SITARA and Spain’s Looking For Perseida placed second and third, and Kazakhstan’s Myrzakhan Zhumakhan won Best Character for It Was Just a Joke.
Qazinform
Added Oct 4, 2026 7:53 am ET
Creative Bloq
On October 4, 2026, Creative Bloq reported on newly published Adobe patents, including a pipeline that turns a text prompt into a spherical intermediate view and then navigable 3D geometry. The piece stresses these are filings—not shipping Firefly features—alongside related ideas for new camera angles from one photo and 3D garment previews.
Creative Bloq
Added Oct 4, 2026 7:53 am ET
Scienmag
Coverage dated October 4, 2026 highlights Amit Raphael Zoran’s open-access AI & Society paper “More than we can make.” It argues generative AI loosens the historical link between sophisticated forms and the long maker’s journey that once produced them, and urges practices that reinvest saved effort into embodied, situated experience rather than endless symbolic recombination.
Scienmag
Added Oct 4, 2026 7:53 am ET
HeyGen
HeyGen’s open-source HyperFrames framework treats web markup as a timeline: an AI agent writes HTML, CSS, and JavaScript, then HyperFrames seeks exact frames for a reliable MP4, MOV, WebM, or GIF render. Projects stay editable folders rather than flattened black boxes, with local CLI, Studio, and hosted cloud render paths documented on HyperFrames’ site.
HeyGen
Added Oct 4, 2026 7:53 am ET
Higgsfield
On October 2, 2026, Higgsfield opened AI Influencer: build a character from tiles or one face photo, then drop them into a curated trending motion or a 3–30 second reference clip. Genjutsu handles motion transfer and object swaps so outfits, products, or locations can change while the take stays intact, and the same flow is free to try in Higgsfield’s ChatGPT extension.
Higgsfield
Added Oct 3, 2026 7:51 pm ET
Tavus
On October 1, 2026, Tavus introduced Griffin, a Human Interaction Model that generates full-scene video and speech in real time while watching and listening. Griffin-Lite is a research preview for trusted testers; Tavus reports 48% of study participants thought a one-minute call was with a real person, and says a wider release waits on further safety work.
Tavus
Added Oct 3, 2026 7:51 pm ET
Opus6
Coverage dated October 1, 2026 highlights opus6.video, a browser tool where Claude Opus 5.5 plans a storyboard and writes each scene as code—canvas, type, timing—then records a 1080p MP4. Edits stay local to a scene instead of regenerating pixels, which suits launch clips, explainers, and showreels without a Claude subscription.
Opus6
Added Oct 3, 2026 7:51 pm ET
Recraft
On October 1, 2026, Recraft said Ideogram 4.5 is coming to Recraft Studio beside Recraft V4.1. The hosted model is built to hold untouched pixels steady across many edits, keep native resolution on large sources, and handle jobs from text and lighting tweaks to product shots and restorations without forcing a full re-generate.
Recraft
Added Oct 3, 2026 7:51 pm ET
Digital Matters
As of October 2, 2026, gemini-2.5-flash-image—the first stable Nano Banana—reached its Gemini API shutdown date. Google’s image guide steers pipelines to Nano Banana 2 (gemini-3.1-flash-image) or Nano Banana 2 Lite (gemini-3.1-flash-lite-image); Digital Matters warns the deprecations page still names a preview ID that already shut down in June.
Digital Matters
Added Oct 3, 2026 7:51 pm ET
EIN Presswire
On October 3, 2026, BeatViz announced AI Director, a conversational workflow that turns a finished track into a planned music video: visual direction, characters, scenes, shot lists, generation, and revision stay tied to the song rather than one-off clips. It sits alongside BeatViz Workflow and a timeline Editor so creators can rework a chorus location or shot density without restarting from scratch.
EIN Presswire
Added Oct 3, 2026 3:57 pm ET
Kotaku
On October 2, 2026, Kotaku covered illustrator Loish (Lois van Baarle) after TikTok commenters claimed a piece was AI-made. She pointed to process videos showing the work was drawn in 2023, plus two decades of public style evolution, and urged people to check profiles and process before accusing artists—especially when AI systems were trained on their labor without consent.
Kotaku
Added Oct 3, 2026 3:57 pm ET
DEVELOP3D
On October 2, 2026, DEVELOP3D reported ClassCAD.ai from AWV Informatik: a self-contained Model Context Protocol layer that lets assistants such as Claude build parametric or direct solid models with a headless NURBS kernel (native or WebAssembly). Users describe parts or paste requirements, then export to conventional CAD; non-commercial use is free, with commercial licenses from 100 CHF.
DEVELOP3D
Added Oct 3, 2026 3:57 pm ET
MIT News
MIT News reported on October 1, 2026 that InstructMesh (MIT CSAIL, Google, and Northeastern) pairs TRELLIS 3D generation with an LLM so users can highlight flawed regions and refine them in plain language before fabrication. In tests, about 80% of TRELLIS recreations of Thingiverse models had structural issues, and novices spotted and fixed fabrication problems about 90% of the time. The work heads to UIST in November.
MIT News
Added Oct 3, 2026 3:57 pm ET
IntraGoals
Coverage dated October 2, 2026 describes Stillwet Art Studio, where models such as Claude Opus 5.5 write code that drives a simulated oil-on-linen canvas—bristles, wet paint, drying, and glazes—instead of calling an image generator. The project has produced dozens of Friedrich- and Hopper-manner paintings across easel rounds with look-and-revise steps; paintings, timelapses, and code are published at stillwet.art.
IntraGoals
Added Oct 3, 2026 3:57 pm ET
Jitter
On October 1, 2026, Jitter said Jitter AI and AI Shaders now run on Anthropic’s Claude Opus 5.5, with stronger motion, transitions, and sequencing from a prompt. Animations land as editable layers on the canvas, artboard variations can sit beside the original, and generations use about 25% fewer credits. Failed image-to-video and background-removal jobs no longer spend credits.
Jitter
Added Oct 3, 2026 11:56 am ET
YouArt
On October 2, 2026, YouArt released Logo Animation AI so teams can turn a PNG or SVG logo—plus a prompt or reference image—into motion for intros, social openers, sites, and email. The browser tool keeps brand colors and type in play, exports MP4 or GIF, and offers free trial credits before a paid plan.
YouArt
Added Oct 3, 2026 11:56 am ET
Parametric Architecture
On October 2, 2026, Parametric Architecture reported FORMAS.AI’s CARTESIAN preview for architecture, interiors, furniture, and product design. Designers start from photos, sketches, briefs, or models; AI agents author editable geometry through a CAD kernel, with Rhino, Blender, and SketchUp links. Desktop access is phased for macOS and Windows, with wider access planned from October 31, 2026.
Parametric Architecture
Added Oct 3, 2026 11:56 am ET
Alia Newswire
Coverage dated October 2, 2026 describes Bake3D, which builds a textured 3D model from a drawing, render, or photo, then helps plan a rig and generate motion for games, film, and cartoons. It aims beyond standard humanoids—animals, creatures, robots, vehicles—and exports GLB/FBX, with an API for agent pipelines. Results still need a human pass before production use.
Alia Newswire
Added Oct 3, 2026 11:55 am ET
Figma
Figma’s help center says the Design Agent and Weave tools move from open beta to general availability on October 6, 2026, after which they draw from the same AI credit pool as features like Figma Make. Both stay free during beta; Weave tools will have set per-tool costs, while agent usage will vary with prompt complexity and attached context.
Figma
Added Oct 3, 2026 11:55 am ET
SLOP TV News
On October 1, 2026, Runway Labs showed Project Continuum, an early research application that treats parts of the desktop as live generated video rather than fixed app windows. The demo names four modes—Portals, Visual Thinking, Responsive Video Interfaces, and Interactive Worlds—powered by Solaris, Runway’s Interface World Model on Gen-4.5. Resizing a portal regenerates the scene; nothing is for sale yet, and Runway published no waitlist, price, or release date.
SLOP TV News
Added Oct 3, 2026 7:56 am ET
fal
Coverage dated October 2, 2026 describes fal’s Recast endpoint (minimax/h3-max/recast), which replaces a person in a short clip with someone from a reference image while aiming to preserve motion, camera path, and original audio. Creators can spin variants without a reshoot; fal bills by output duration (about $0.30/sec at 768p and $0.45/sec at 1080p on the model page). Teams still need clear consent and likeness rights before shipping results.
fal
Added Oct 3, 2026 7:56 am ET
EIN Presswire
On October 1, 2026, Boss Wallah Media said its AI Lab shipped AI Studio, an internal production layer that routes jobs across models such as Kling, Seedance, Flux, GPT Image, and Grok image-to-video under shared brand and spend controls. The company claims roughly 70–80% lower AI production cost versus its prior third-party setup and capacity for more than 1,000 videos a month. The tool is internal today, with a later plan to open it to outside studios.
EIN Presswire
Added Oct 3, 2026 7:56 am ET
Deadline
In late September 2026, Jeff Synthesized’s AI-produced short The Gifted won the inaugural Future Vision XPRIZE from more than 2,500 entries. Deadline reports a $2.5 million equity package toward a feature, plus $100,000 for screenplay development, with Range Media Partners and Google’s 100 ZEROS initiative supporting production. The story follows a boy who rebuilds his late mother from her texts, voice, and diary; it also took Snap’s SPECS Audience Choice top prize.
Deadline
Added Oct 3, 2026 7:56 am ET
Creati.ai
On October 3, 2026, Creati.ai covered University of Maryland and AWS research on LEGO-Anything, which asks coding agents to write and revise Blender programs that rebuild a 3D scene from one photo. On LEGO-Bench, the strongest reported setup reached about 53% indoor and 40% outdoor reconstruction accuracy, while agents’ judgments of which version was better scored near chance. A measurement-focused LEGO-Plugin improved results without retraining the models.
Creati.ai
Added Oct 3, 2026 7:56 am ET
TechCrunch
On October 2, 2026, TechCrunch reported that Napster co-founder Sean Parker and CEO Prem Akkaraju are rebuilding Stability AI as a music-first AI toolmaker for professionals. After a late-August raise that included Sony, Warner, and Universal—who also licensed catalogs for training—the company has shipped three audio models and music-editing software that can generate instrumental tracks or short snippets from text. Parker says a coming update will let users hum a melody or beatbox a drum pattern to steer the result.
TechCrunch
Added Oct 2, 2026 7:52 pm ET
mnml.ai
On October 1, 2026, mnml.ai launched an MCP connector so architects can start, refine, and upscale client-ready renders from Claude, ChatGPT (developer mode), or other OAuth-capable assistants without leaving the chat. Renders run on the user’s mnml credits and engines, save to their Studio library, and support sketches, photos, model views, or text briefs. Free accounts can connect; the assistant asks before spending credits, and mnml says it does not train models on customer content.
mnml.ai
Added Oct 2, 2026 7:52 pm ET
Pressat
On October 2, 2026, Magic Hour said its AI Video Editor now offers a free browser workflow (no signup for basic use), image-reference guidance in compatible modes, and programmatic API access. Editors upload a clip, describe visual changes in natural language—remove, replace, restyle, enhance—and can pick models and resolutions. The company positions it for creators, agencies, and developers who want generative edits without rebuilding a full timeline.
Pressat
Added Oct 2, 2026 7:52 pm ET
ABNewswire
On October 1, 2026, InfoseekAI launched an MCP-native Seedance video agent that turns plain-language direction—plus optional image, video, or audio references—into short clips returned in the chat. The agent guides subject, action, setting, style, and sound, supports Seedance 2.5, and saves shareable results to an Infoseek library. Pricing is pay-as-you-go with no subscription; creators can also run the same agent from Infoseek.ai.
ABNewswire
Added Oct 2, 2026 7:52 pm ET
The Verge
On October 1–2, 2026, Suno opened Speech in public beta on web and mobile: generate spoken audio from a prompt or a custom script, optionally scored with original background music in one track. The Verge notes Simple and Advanced modes, controls for voice gender, style, and variety, a roughly eight-minute cap, and a toggle for clean speech without music. Suno calls the beta imperfect—accents and dramatic pauses can wander—and says it will improve from community use.
The Verge
Added Oct 2, 2026 3:59 pm ET
SLOP TV News
SLOP TV News reported on October 2, 2026 that xAI’s grok-imagine-video-1.5-lite is live on the xAI API and fal.ai for text-to-video and image-to-video with native audio. Docs list $0.02 per second at 480p, $0.03 at 720p, and $0.14 at 1080p, plus $0.01 per input image—cheaper than the full 1.5 tier at every matching resolution. Regions include us-east-1 and us-west-2, with a 10 requests-per-second cap and batch API support.
SLOP TV News
Added Oct 2, 2026 3:59 pm ET
EIN Presswire
On October 1, 2026, ManualFig announced an AI-native browser workflow for assembly guides, installation steps, and product manuals. Creators start from reference photos or CAD screenshots plus written instructions, then revise figures in plain language—viewpoint, arrows, labels—and export PNG or SVG (or a ZIP of project assets). Task pages cover line art, exploded views, work instructions, troubleshooting, and safety figures, while the documentation team still validates procedures and compliance.
EIN Presswire
Added Oct 2, 2026 3:59 pm ET
Ubergizmo
Ubergizmo’s October 1, 2026 review covers Photon Studio from Tenzen Studio: a free desktop image editor for Windows, macOS, and Linux with layers, masks, PSD support, and Photoshop-like shortcuts, and no subscription or AI credit wall. Subject selection, background removal, and assistive tools run locally so images are not uploaded for cloud training. Early notes include occasional untranslated menus and heavy-canvas slowdowns; downloads are on tenzen.studio and the Microsoft Store.
Ubergizmo
Added Oct 2, 2026 3:59 pm ET
Tubefilter
On October 2, 2026, Tubefilter highlighted Firefly’s Generate Music, Generate Speech, and Generate Sound Effects as commercially safe audio tools creators can use inside Adobe’s generative studio alongside image and video. Adobe’s Meg Donovan describes Generate Music as original tracks tuned to a clip’s length and mood without licensing scavenger hunts. The same piece notes free limited daily access to Firefly AI Assistant, with paid plans from about $9.99/month for 2,000+ generative credits.
Tubefilter
Added Oct 2, 2026 3:59 pm ET
TechCrunch
On October 2, 2026, TechCrunch reported that Pope Leo XIV posted on X arguing it is urgent to separate human art from machine-made images. He called the gap ontological, not only aesthetic: algorithms assemble from millions of other people’s pictures and lack a human spark. The story notes his earlier encyclical on dignity in an automated world, and that Anthropic has lobbied the Vatican on non-human consciousness without much success.
TechCrunch
Added Oct 2, 2026 11:55 am ET
Midjourney
Midjourney’s October 1, 2026 alpha changelog (posted October 2) focuses on cleanup after last week’s style previews, larger Create images, and per-folder defaults. Edits keep the base image’s aspect ratio, style references collapse into one pill with thumbnails, Organize selection no longer accidental-drags images, and deep history no longer crashes large libraries. Collaborative sharing tools and persistent edit history are flagged for next week.
Midjourney
Added Oct 2, 2026 11:55 am ET
Cartoon Brew
Ahead of LightBox Expo on October 23–25, 2026 in Pasadena, Cartoon Brew interviewed co-founders Bobby Chiu and Jim Demonakos about widening a years-old Artist Alley rule into an event-wide stance. They equate generative AI work with bootleg merch and say human craft is the point of the show. The official policy still allows minor helpers like spell check, upscalers, basic filters, and accessibility tools when the core creative work stays human.
Cartoon Brew
Added Oct 2, 2026 11:55 am ET
The Keyword
On October 1, 2026, OpenAI put a Try on button on clothing and accessory results in ChatGPT worldwide, The Keyword reports. Shoppers upload a selfie and ChatGPT Images 2.5 renders the garment on them; Favorites saves products and try-ons into the ChatGPT Library. OpenAI warns the images may not match fit or appearance exactly, and product results are selected by ChatGPT rather than sold as ads.
The Keyword
Added Oct 2, 2026 11:55 am ET
Creatify Labs
On October 2, 2026, Creatify Labs launched Boreal-H3, a MiniMax H3 model post-trained for advertising: readable labels, stable packaging, and creators who stay the same person across a clip. Creatify says it scored about a third higher than base MiniMax H3 on its Ads Quality Index and costs $0.04 per second at 768p. It is live in the Model Playground and API with text-, image-, and multi-reference video modes.
Creatify Labs
Added Oct 2, 2026 11:55 am ET
Comfy
On October 1, 2026, Comfy announced Comfy Agent for Comfy Cloud (Desktop coming later): describe a goal and the agent plans, builds, runs, and fixes ComfyUI workflows on the same canvas you edit. It knows current models and extensions, supports public or private skills, and can run up to five parallel chats. Early creator examples cover character animation and VFX masking; the beta uses existing Comfy Credits with free starter chats.
Comfy
Added Oct 2, 2026 7:51 am ET
THE DECODER
Black Forest Labs has made FLUX 3 Image available through its API and docs: one endpoint for text-to-image and instruction edits, with bounding-box layouts, up to ten reference images, and output up to 4K. THE DECODER reported the October 2, 2026 launch, noting multi-step edits that aim to leave untouched regions alone, a 50% API discount through October 8, commercial weights for self-hosting, and open weights expected later.
THE DECODER
Added Oct 2, 2026 7:51 am ET
Creative AI News
Ideogram 4.5 launched on September 30, 2026 as an image model aimed at repeated edits without color drift or artifact buildup. Creative AI News reports Precise Edit returns the source dimensions and copies unchanged pixels, with Generate plus Edit for freer changes; it is live in Ideogram’s app and API and on partners including fal, Pika, Krea, Picsart, and ComfyUI, with fal edit pricing from about $0.008 to $0.22 per image. Open weights are promised but not yet dated.
Creative AI News
Added Oct 2, 2026 7:51 am ET
Luma
On October 1, 2026, Luma launched Variants for any Luma user: upload one approved static ad, pick placements and languages, and get resized, translated versions while logo, headline, and CTA stay intact. Five formats ship in the first release—Story 9:16, Portrait 4:5, Square 1:1, Medium Rectangle 6:5, and Widescreen 16:9—with previews before download or export to Luma Canvas. The company frames it for agency and brand multichannel rollouts.
Luma
Added Oct 2, 2026 7:51 am ET
Shopify
On October 1, 2026, Shopify introduced Canvas, a visual workspace rolling out over coming days where merchants pan across every store page and collaborate with Sidekick, Shopify’s AI agent. Sidekick proposes design directions, edits real theme code, validates changes, and reviews screenshots while merchants click elements or chat feedback in real time. The company positions it so bespoke store design no longer requires coding or a heavy settings panel; it does not yet replace the existing editor.
Shopify
Added Oct 2, 2026 7:51 am ET
invideo
On September 30, 2026, invideo launched invideo MCP, a Model Context Protocol server that connects agents such as Claude, ChatGPT, Cursor, and Gemini CLI to invideo Editor. Agents can drive timeline edits, color, audio, and image/video/audio generation in a browser workspace while creators keep a full multitrack editor, version control, and export paths to Premiere, Final Cut, or DaVinci Resolve. The manual editor is free; AI generation uses invideo credits.
invideo
Added Oct 1, 2026 7:52 pm ET
fal
Pixelcut’s looping-video model is live on fal as pixelcut/looping-video, turning one still product or lifestyle photo into a seamless 5–15 second MP4 with a locked camera. Creators can choose subtle ambient motion or a full 360° spin while labels, logos, and composition stay intact; optional prompts guide motion, and output goes up to 1080p. Pricing on fal is per second of video by resolution, aimed at catalog, fashion, and social loops without a reshoot.
fal
Added Oct 1, 2026 7:52 pm ET
PoYo
PoYo shipped PoYo Agent on September 22, 2026: a conversational workspace where you describe a full creative job, attach references, and let the agent pick image, video, or 3D models—or set your own preferences. Projects keep style notes, decisions, and documents in memory, and reusable Skills encode repeatable instructions such as storyboard or product-shot checklists. Typical flows start with a still, then refine and animate it into a short clip without leaving the chat.
PoYo
Added Oct 1, 2026 7:52 pm ET
TechNode
On September 22, 2026, TechNode reported that SenseTime released the production SenseNova U1 Pro image model through its Raccoon app and SenseNova API. The company says the model uses an interleaved text-image chain of thought for structured layouts and readable text, with output up to 8K and custom aspect ratios for posters, infographics, and large-format work. Availability and pricing outside SenseTime’s announced channels were not detailed in the report.
TechNode
Added Oct 1, 2026 7:52 pm ET
Adobe
On September 8, 2026, Adobe announced Premiere’s Generative Media Tool, which generates editable video and sound effects in the timeline using Firefly plus partner models including Google Veo, Kling, Runway, and Luma. New beta audio tools cover enhance, crosstalk separation, and auto ducking, while After Effects AI Assistant enters public beta to reorganize projects, repair issues, and build expressions from plain language. Color Mode in Premiere also gains Auto Color and new style modules.
Adobe
Added Oct 1, 2026 7:52 pm ET
HeyGen
On September 30, 2026, HeyGen launched HeyGen Video, its first general-purpose video model that builds cinematic clips from text, an image, or a reference video, with dialogue, ambience, and effects in the same pass. The company says the model ranked first in its blind arena tests against Seedance and Kling peers, ships via API (plus OpenRouter, Runware, and ComfyUI), and is on a 50% launch price through October.
HeyGen
Added Oct 1, 2026 4:11 pm ET
Search Engine Watch
Manus shipped version 2.0 on September 28, 2026, with a Video Editor that puts generated clips, images, text, motion graphics, and audio on a timeline so creators can change music or shots without regenerating the whole piece. Search Engine Watch tested a 30-second promo: the first cut took about 21 minutes, and three follow-up edits finished in under four. The update also adds Cascade agents, event-triggered Automations, and optional Cloud Computer plans.
Search Engine Watch
Added Oct 1, 2026 4:11 pm ET
TechCrunch
At Made On YouTube on September 23, 2026, YouTube showed a chat-style editor that takes natural-language directions to pick takes, add music and transitions, overlay text, and trim pauses on long- and short-form footage. Creators can jump back to a normal timeline anytime; TechCrunch notes the feature is aimed at Shorts and the YouTube Create app, with availability planned for early 2027.
TechCrunch
Added Oct 1, 2026 4:11 pm ET
Microsoft
Microsoft unveiled Ink Canvas on September 23, 2026: a pen-first Windows sketchpad for handwriting, markup, AI image generation, and turning canvas notes into text with Copilot. It ships pre-installed on the new Surface Pro 12-inch and hits the Microsoft Store for other supported devices on October 13, 2026; basic inking works without Microsoft 365, while Copilot features need a subscription.
Microsoft
Added Oct 1, 2026 4:11 pm ET
PR Newswire
On October 1, 2026, MixVio AI announced a single creative workspace spanning 14 model families and 26 video, image, and audio generation routes under one account and credit system. Selected video workflows accept up to 50 multimodal references and can render clips up to 30 seconds at 1080p with synced audio, while an image can move into edit or image-to-video without repeated downloads and uploads.
PR Newswire
Added Oct 1, 2026 4:10 pm ET
TechCrunch
On September 30, 2026, Instagram brought Edits Assistant into its Edits video app: a chat tool that reads your account metrics, comments, and trends, then explains what performed and proposes hooks, scripts, captions, and audio ideas. TechCrunch notes Meta frames it as analysis rather than doing the edit for you, with free daily prompt limits and more access via Meta One.
TechCrunch
Added Oct 1, 2026 8:02 am ET
Animation Magazine
On September 30, 2026, Studio Jadu opened as a creator-first animation studio backed by Khosla Ventures and Shakti VC. Its AIDA workbench helps indie artists turn scripts into animation while they keep creative control, plus support for distribution, fandom, and revenue share. Animation Magazine says the Pioneer Program is open, with early shows including Fred Graver’s The Future Was Awesome.
Animation Magazine
Added Oct 1, 2026 8:02 am ET
Trumbone
On October 1, 2026, Trumbone released Indigo (v1.1) for every plan, led by Scene DNA that pulls recurring people, props, and places from a script and reuses the same reference stills across later shots. The update also adds Photo Looks with 24 treatments and Property Tours built only from listing photos, plus stronger lip-sync and scene consistency for music-video workflows.
Trumbone
Added Oct 1, 2026 8:02 am ET
Aju Press
On September 28, 2026, Scatter Lab launched Visual Novel Mode on its Zeta AI character platform, so conversations update character expressions, actions, and backgrounds as the story moves. Aju Press reports real-time image generation for new characters, choice and dice events, outfit application from creator shops, and a path from text chat into visual storytelling.
Aju Press
Added Oct 1, 2026 8:02 am ET
Creative Bloq
Creative Bloq reports on La Triste Figura, Guillermo Miranda and Javier de la Chica’s AI-centered Don Quixote short that premiered in Madrid on September 23, 2026. About 120 shots came from roughly 1,200 generations, mostly Seedance 2.5 and Seedream 5 Pro over hand storyboards and design sheets, with human voices and hard limits showing up in long comic dialogue and tiny continuity fixes.
Creative Bloq
Added Oct 1, 2026 8:01 am ET
Runway
On September 30, 2026, Runway launched Runway Ads, which connects a brand kit and ad account, generates video and image variants, and publishes human-approved creatives to Meta, Google, and TikTok. It then reads performance back from those platforms and produces the next round around what earned spend, with localization, brand checks, and human approval on by default.
Runway
Added Sep 30, 2026 7:58 pm ET
Business Wire
Utopai Studios introduced PAI, a production-intelligence platform that keeps screenplay, characters, locations, and shots connected while filmmakers plan, generate, and revise. Utopai X is the video model inside PAI; Business Wire says it ranks No. 2 on Artificial Analysis’ Text-to-Video Leaderboard With Audio and is the highest-ranked U.S.-based model on that board.
Business Wire
Added Sep 30, 2026 7:58 pm ET
MediaBrief
At FICCI FRAMES 2026 on September 29, Eloelo Group opened Dolphin AI, an in-house video studio now available to creators, brands, and filmmakers. MediaBrief reports multi-shot planning that aims to keep characters, voices, wardrobes, and locations consistent across longer pieces, plus Film Studio script breakdown and MCP access for Claude, ChatGPT, Gemini, and Cursor.
MediaBrief
Added Sep 30, 2026 7:58 pm ET
EIN Presswire
On September 28, 2026, HeyVigo launched Infinite Canvas, a node-based workspace where image, video, text, and audio sit on one surface with links that preserve how references become outputs. EIN Presswire says creators can search, group, and reuse assets, keep confirmation on for AI-proposed canvas edits, and use undo and saved-state markers while building ads, storyboards, or short dramas.
EIN Presswire
Added Sep 30, 2026 7:58 pm ET
LemonSlice
On September 23, 2026, LemonSlice released Character World Model-1, a real-time interactive avatar model that generates face, body, hands, and scene from audio and prompts as a conversation runs. The company says an emotion engine steers actions and expressions, any character can be started from one image without fine-tuning, and the web demo is free with API access for Ultra and Enterprise plans.
LemonSlice
Added Sep 30, 2026 7:58 pm ET
CellCog
On September 28, 2026, Kling said Kling 4.0 Flash is live for Ultra Yearly subscribers while the full Kling 4.0 is due in October. The announcement pitches native 30-second clips, up to 10 keyframes, up to 15 multi-modal references, stereo audio with stronger lip-sync, and later 4K with 10-bit HDR.
CellCog
Added Sep 30, 2026 3:56 pm ET
EIN Presswire
On September 29, 2026, Bandicam Company said Bandicam AI Studio is live, built around an AI Director that turns a short idea into a storyboarded multi-scene video. The suite also lists AI subtitles, lip sync, motion control, upscaling, and MCP hooks for clients such as Claude, Codex, Cursor, and ChatGPT, with free trial credits for new users.
EIN Presswire
Added Sep 30, 2026 3:56 pm ET
fal
fal’s Seedance 2.0 4K page says ByteDance’s highest-fidelity Seedance 2.0 tier is live on fal.ai for text, image, audio, and video inputs, rendering short cinematic clips in up to 4K with native synced audio. It positions the 4K tier for final renders and hero shots, with a faster Mini option for everyday drafts.
fal
Added Sep 30, 2026 3:56 pm ET
Google
On September 24, 2026, Google introduced Gemini 3.8 Live with Live Avatar in Gemini Enterprise, pairing live dialogue with low-latency streaming video so an animated persona can listen, speak, and lip-sync. Enterprises can use preset avatars or customize from a reference image, with SynthID watermarking on the generated audio and video.
Google
Added Sep 30, 2026 3:55 pm ET
Newsworthy.ai
A September 16, 2026 Newsworthy.ai note says Canvora’s visual engine now plugs into Claude and ChatGPT over MCP, so an agent can turn briefs, docs, and links into branded deliverables across 100+ formats and 150+ languages. The remote server exposes generation, edit, localize, and export tools, with a confirm step before credits are spent.
Newsworthy.ai
Added Sep 30, 2026 3:55 pm ET
ACCESS Newswire
On September 30, 2026, Maxon shipped Model Context Protocol support in Cinema 4D 2026.4 so artists can ask Claude or ChatGPT to handle repetitive 3D chores—scene organization, material variations, multipass prep, basic modeling, UV, and rigging—through Cinema 4D’s own tools. The server stays off by default, runs locally with tool permissions and an audit log, and leaves native, undoable scene data.
ACCESS Newswire
Added Sep 30, 2026 11:54 am ET
Adobe
Adobe’s September 24, 2026 post says Adobe in Gemini is rolling out on every Gemini plan, letting people enhance photos and build marketing assets through chat with tools from Photoshop, Lightroom, Express, and Firefly. The same update unifies Acrobat with Adobe’s creative tools in Claude and adds interactive editors for PDFs and Express designs.
Adobe
Added Sep 30, 2026 11:54 am ET
EIN Presswire
Sekondo, a plugin by Spanish designer Jose Manuel Fernández, went on sale September 30, 2026. It connects Claude or ChatGPT to Photoshop, Illustrator, InDesign, After Effects, Premiere Pro, and Lightroom Classic so spoken or typed requests run inside open documents while layers, text, and styles stay editable. It runs locally on Mac and Windows; reserving a seat is free until delivery.
EIN Presswire
Added Sep 30, 2026 11:54 am ET
PixAI
On September 16, 2026, PixAI made Tsubaki.3 available to all creators after early access. The model targets ACG work with clearer multilingual text in-image, multi-panel manga, stronger multi-character scenes, up to three smart references, Repaint, and a Color Palette, plus a Tsubaki Video line for camera-directed animation. A launch offer discounts video credits through October 20.
PixAI
Added Sep 30, 2026 11:54 am ET
LBB
Canva announced ProSuite, tying Affinity design, Cavalry motion, Flourish data viz, and Leonardo generative variations into one professional path that can hand off into Canva for scale. LBB reports more than 100 new features, live-editable motion assets in Canva, and brand-kit import from Affinity, with AI framed as help for repetitive batch and format work rather than replacing craft.
LBB
Added Sep 30, 2026 11:54 am ET
Creative Bloq
Creative Bloq reports that Adobe added Relight to Photoshop beta on September 28–30, 2026: a generative layer that strips existing lighting and lets you place up to three interactive lights with intensity, color, and diffusion controls. Early reactions are split—some see a fast base for retouching, others say portraits drift toward a smoothed, game-like look—and each final bake costs Firefly credits.
Creative Bloq
Added Sep 30, 2026 7:55 am ET
Pika
On September 17, 2026, Pika unveiled a rebuilt creative platform with 25+ task-focused apps—Video Studio, Relight Media, Product Shot, Camera Director, and more—backed by models including Seedance, GPT Image, MiniMax H3, and Pika’s own. Apps pick a model for you unless you choose one, and the company says plan tiers differ by credits and concurrency rather than gated tools.
Pika
Added Sep 30, 2026 7:55 am ET
Google
Google’s September 23, 2026 Workspace note says anyone with a Google account can generate 1080p clips in Google Vids using Gemini Omni 1.1 Flash at no cost. Creators can extend scenes, set exact clip lengths, start from templates, and every Omni clip carries a SynthID watermark for transparency.
Google
Added Sep 30, 2026 7:55 am ET
Games Press
On September 23, 2026, Ludo.ai announced an AI Texture Generator that turns a surface prompt into seamless, tileable textures in more than 30 styles, with optional motion prompts for looping water, lava, clouds, and light. Exports include sprite sheets, GIFs, or frames, and the tool plugs into Ludo’s API and MCP workflow for indie and small-studio pipelines.
Games Press
Added Sep 30, 2026 7:55 am ET
Runway
Runway said on September 29, 2026 that Runway Creative is available in the OpenAI Marketplace as a launch partner, so eligible enterprise customers can apply OpenAI spend commitments toward image, video, audio, and editing models in one workspace. The listing covers Gen-4.5, Seedance 2.5, GPT Image 2.5, ElevenLabs V4, and agent-style brief-to-video workflows.
Runway
Added Sep 30, 2026 7:55 am ET
Google Research
On September 29, 2026, Google Research described Diffusion Controller, a small add-on network that guides text-to-image models toward better prompt matches while the base model stays frozen. It can attach even when the host model is closed, and Google says the white-box version beat its baseline on about nine in ten human preference comparisons.
Google Research
Added Sep 29, 2026 7:55 pm ET
Sogni
Sogni’s September 27, 2026 note says MiniMax H3 can now land on as many as eight stills at exact times between the first and last frame, not only the endpoints. The feature works on Image to Video, Sound to Video, and Reference to Video, and Sound to Video keeps lip-sync through the cuts.
Sogni
Added Sep 29, 2026 7:55 pm ET
Agile Brand Guide
Adobe said on September 29, 2026, at OpenAI DevDay, that its ChatGPT plugin now opens Photoshop- and Express-powered editors beside the chat, plus an Acrobat PDF panel. You can prompt a change, then finish with sliders and brushes; edits feed back into the conversation so the next ask keeps that context.
Agile Brand Guide
Added Sep 29, 2026 7:55 pm ET
Midjourney
In a September 24, 2026 alpha note, Midjourney says Live previews can show your current prompt across liked and featured styles before you commit. The edit model’s inpainting and outpainting now change only selected pixels for cleaner repeat edits, and --tile on V8.1/8.2 is meant to hide seams between repeats.
Midjourney
Added Sep 29, 2026 7:55 pm ET
PetaPixel
Imagen began staging Intelligent Studio to existing customers on September 29, 2026, with a built-in manual editor, AI culling and filters, and an Assistant that runs multi-step edits from plain language. Photographers can move between cull and edit in any order, and pay-as-you-go billing now follows exported photos instead of each AI click.
PetaPixel
Added Sep 29, 2026 7:55 pm ET
Google
On September 29, 2026, Google Arts & Culture Lab and LAS Art Foundation posted the outcomes of an online artist residency. Xin Liu’s “nom nom” explores swarm-like collective intelligence with neural cellular automata research, and Eunjo Lee’s “In Your Eyes, Before My Own” turns hand sketches into 3D scenes with Google AI tools including Nano Banana and Veo.
Google
Added Sep 29, 2026 3:53 pm ET
The Gauntlet
A September 28, 2026 Gauntlet feature revisits Calgary Public Library’s Artificial Intelligence Collaborative Artist residency after community backlash earlier this year. Resident Paul Freeman, who finished the June–September term, describes his Stable Diffusion practice as treating AI as a medium rather than a paintbrush, while critics still question energy use, consent, and replacing human artists.
The Gauntlet
Added Sep 29, 2026 3:53 pm ET
WPR
Wisconsin Public Radio reports that Beeple’s interactive “Diffuse Control” installation opened at the Madison Museum of Contemporary Art around September 25, 2026, after more than 900 people signed an open letter against it. Viewers can drive changes across twelve screens fed from collection selections, while critics object to diffusion training practices and consent for artists whose work enters the piece.
WPR
Added Sep 29, 2026 3:53 pm ET
Midjourney
Midjourney’s September 23–24, 2026 alpha changelog adds live style previews for the current prompt, savable default parameters, a full-width masonry Create feed, and Korean language support for everyone on midjourney.com. Editing history navigation, project organize tools, and several prompt-bar fixes ship in the same pass.
Midjourney
Added Sep 29, 2026 3:53 pm ET
MLive
In a September 21, 2026 report, ArtPrize organizers pushed back on social-media claims that fully AI-made entries can win prizes. They say generative tools are allowed only as part of the longstanding New Media category, humans alone can receive awards, fully AI-created work is barred, and guidelines will be reviewed with artists after the festival.
MLive
Added Sep 29, 2026 3:53 pm ET
Creative Bloq
CorelDRAW Graphics Suite’s September 2026 update adds Text-to-Vector, which builds curves and nodes from a prompt or a reference image so the result can be edited like any other drawing. Creative Bloq says the generation runs in the cloud, in three models from simple icons to more complex variations of existing artwork.
Creative Bloq
Added Sep 29, 2026 1:56 pm ET
Runway
On September 23, 2026, Runway released a plugin for DaVinci Resolve Studio that generates and restyles footage from inside the editor and places the result on the timeline. It requires Resolve Studio 19 or later, and generations draw on an existing paid Runway plan.
Runway
Added Sep 29, 2026 1:56 pm ET
Adobe Community
Adobe’s September 18, 2026 Firefly note adds one-step background removal for a whole video clip, more aspect ratios for GPT Image 2, and reference images for Kling 3.0 Omni. It also lengthens image and video prompts and says the assistant’s tone was pulled back so the focus stays on the work.
Adobe Community
Added Sep 29, 2026 1:56 pm ET
OpenAI
OpenAI’s help center says the Sora web and app experiences closed on April 26, 2026, and the Sora API was discontinued on September 24, 2026. Makers are told to export what they created; after any final export window, OpenAI says it will delete data tied to Sora use.
OpenAI
Added Sep 29, 2026 1:56 pm ET
OpenAI
On September 8, 2026, OpenAI introduced ChatGPT Images 2.5, describing sharper detail, more reliable multi-step edits, and latency cut by up to half compared with Images 2.0. In ChatGPT you can sketch a reference, start from templates, and leave comments on an image for a tighter edit.
OpenAI
Added Sep 29, 2026 1:56 pm ET