Google DeepMind and A24 Strike $75 Million Deal for AI Filmmaking Tools, SAG-AFTRA Ratifies New Deal With Expanded AI Protections, No Fakes Act risks sweeping up video game industry in bid to curb AI deepfakes
Google DeepMind and A24 Strike $75 Million Deal for AI Filmmaking Tools
Google DeepMind has partnered with A24 in a $75 million deal to develop AI filmmaking tools for the indie studio’s production and distribution teams. The partnership will support A24 Labs, a 20-person unit building new applications for filmmakers, including a tool for AI-generated storyboards, while also feeding into Google’s wider AI ecosystem. A24, the independent studio behind some of the most culturally influential arthouse and genre films of recent years, has helped bring a more stylized kind of cinema into the mainstream, which makes the deal a notable bridge between Hollywood taste-making and Silicon Valley’s push into generative AI. The agreement marks Google’s first investment in a Hollywood studio and extends its reach beyond YouTube and its Veo video model. It also stops short of allowing Google to train models on A24’s film and TV catalog, suggesting a narrower focus on creative workflow tools than on direct use of the studio’s intellectual property. Google said the collaboration’s technical goals and creative milestones will evolve, pointing to a longer-term research and development relationship across multiple projects.
SAG-AFTRA Ratifies New Deal With Expanded AI Protections
SAG-AFTRA members have overwhelmingly ratified a new four-year contract with the major studios, giving the union a landslide win and locking in stronger protections around artificial intelligence, pay, and benefits through 2030. The agreement was approved by more than 90 percent of voting members and was presented as one of the most consequential labor deals Hollywood has seen in years. The contract builds on the union’s earlier AI rules by tightening limits on synthetic performers and digital replicas. Under the new terms, producers can use AI-generated performances only when they provide “significant additional value” beyond a live actor or that actor’s digital capture, and the union says the rules are meant to keep human performance at the center of production.
The deal also includes broader gains beyond AI. It raises wages, improves benefits, and merges the union’s two pension plans, while giving members a more secure framework for work in an industry that is rapidly changing under streaming and generative technology. The vote reflects how much AI has become a defining bargaining issue in Hollywood since the strikes. SAG-AFTRA leaders argued that the final contract draws a harder line against unauthorized digital replacement, while still leaving room for limited uses of AI when the technology adds something genuinely new to a project.
How Obsidian Studio Is Using AI to Transform Filmmaking
Obsidian Studio is trying to reframe the way Hollywood thinks about AI, pitching the technology as a tool that can speed up production without replacing human creativity. The company’s founders, Wes Walker and Louis Gheysens, say their approach is built around collaboration between directors, artists, and machine-generated workflows. In practice, that means using AI across story development, previsualization, and other parts of production, while still keeping people in the loop at key creative stages. Obsidian has also drawn attention for pairing its AI pipeline with experienced industry artists and for forming a creative partnership with Imagine Entertainment. Their message is, AI can help filmmakers move faster, iterate sooner, and explore more ideas, but the company says the process still starts with human judgment. That balance between automation and artistry is at the center of Obsidian’s attempt to position itself as a future-facing studio rather than a replacement for traditional filmmaking.
Publishers sue Meta over AI training, setting up high-stakes copyright battle
A coalition of major publishers and author Scott Turow has filed a class action lawsuit against Meta in Manhattan federal court, accusing the company of using millions of copyrighted books and academic works without permission to train its Llama AI model. The case, brought in 2026 by firms including Hachette, Macmillan, and Elsevier, claims Meta relied on pirated libraries and scraped content, framing the practice as a large-scale copyright breach. Meta rejects the allegations, arguing AI training constitutes fair use, a defence now central to a growing wave of legal battles between tech companies and creators. The lawsuit lands amid intensifying tensions between the AI sector and creative industries, from publishing to Hollywood, as generative tools become embedded in production. At stake is how future models are built and whether companies will be forced into costly licensing deals for training data. The case also adds to mounting scrutiny of Meta, as competition with OpenAI and Google accelerates and pressure grows for clearer legal boundaries around AI development.
No Fakes Act risks sweeping up video game industry in bid to curb AI deepfakes
Congress is fast-tracking a bipartisan bill aimed at curbing AI deepfakes, but the proposed law could inadvertently disrupt core practices in the video game industry. The No Fakes Act, reintroduced in both chambers in 2026, seeks to protect individuals’ voice and likeness from unauthorised digital replication. Its broad definition of a “digital replica,” covering highly realistic, computer-generated representations of identifiable people, whether altered or entirely synthetic, risks sweeping in standard game development techniques, from photorealistic athletes to procedurally generated background characters.
Industry body, The Entertainment Software Association, has warned lawmakers, including Senators Chuck Grassley and Dick Durbin, that the bill fails to distinguish between harmful deepfakes and legitimate creative uses. It argues this could expose studios to costly litigation, even in cases of coincidental resemblance, and create uncertainty around widely used tools such as avatar builders, motion capture systems, and character engines. The stakes are particularly high for sports franchises and other titles built on licensed likenesses, where existing agreements could be destabilised without clear exemptions. Developers of user-generated content systems may also face liability if players create characters resembling real individuals. While the legislation responds to genuine harms posed by AI-generated deepfakes, critics say its current scope reflects a lack of technical nuance. Without refinement, it could chill innovation in one of the US’s most commercially and culturally significant creative sectors, prompting studios to reassess licensing, pipelines, and legal exposure as the bill advances.
China’s AI Actors Spark Backlash Over Celebrity Lookalikes and Portrait Rights
China’s use of AI-generated actors has sparked backlash after production company Yaoke Media introduced two virtual performers, Qin Lingyue and Lin Xiyan, for a short drama. The pair were promoted across Weibo, Douyin, and Xiaohongshu, but the rollout quickly drew criticism from viewers who said the faces resembled real celebrities, reviving concerns over portrait rights and consent. A company rep told local media the faces were created “without copying or using the facial features of any real individual,” but regulators and industry bodies have continued to stress lawful training-data sourcing and stronger protection for performers’ portrait and voice rights, as the OECD.ai incident monitor logs the episode as a dispute over intellectual property and image rights.
The controversy lands at a moment when AI actors are spreading rapidly through China’s micro-drama and animated short-drama sector. SixthTone, citing DataEye data, reported that AI performers appeared in almost 40% of the top 100 animated short dramas in January 2026, up from less than 10% a year earlier, while the wider micro-drama market reached 100 billion yuan, or about $14.5 billion, last year.
