Google DeepMind and A24 Plan AI Tools For Filmakers

Google DeepMind, Google’s AI research lab, plans to build filmmaker-shaped AI tools with independent film studio A24, with creator feedback and limits on A24 library data access at the center.

Google made an investment in A24 with a reported amount of about $75 million,.

The multiyear deal’s reported non-exclusive structure and lack of direct A24 library data access keep it separate from an acquisition, a finished product launch, or a catalog-training arrangement.

How the A24 Research Deal Is Supposed to Work

DeepMind researchers are expected to work with A24 and its filmmakers across multiple projects, with A24 artists providing feedback and guidance. Eli Collins, VP Product, Google DeepMind, framed the roadmap as evolving over time rather than starting with fixed technical outputs or creative milestones.

Scott Belsky, A24 partner and technology division lead, positioned the collaboration around creative control and support for risk-taking, not replacement. A24 Labs, the studio’s technology arm, is expected to develop AI-generated storyboards, planning images used before production, as an initial application.

Through a deep research and development collaboration, Google DeepMind and A24 could help artists explore ideas without turning every project into a data pipeline. Bringing A24 artists, including Backrooms director Kane Parsons, into the process would make the feedback loop concrete because the tools would touch real creative workflows rather than remain a lab demo.

A storyboard assistant may save time for directors, designers, and producers before a shoot begins. Broader use in editing, lighting, color, visual-effects prep, or distribution planning would move the tools closer to jobs that crews already defend as skilled creative labor.

AI Video Rivals and Hollywood Labor Pressure

AI video tools already sit outside the A24 project. Kling AI, Pika Labs’ shipping AI video tool, Google Veo, Google DeepMind’s video-generation model, and Runway’s AI video-editing models show how crowded the field has become. A24’s project is narrower: studio workflow research instead of a general consumer video generator.

Hollywood’s labor debate gives the partnership its harder test. Storyboards, previsualization, or previs, visual-effects prep, color, lighting, editing, and distribution planning are all areas where faster AI workflows could help production teams or squeeze craft roles.

A 2023 actors’ agreement put consent and compensation for AI use into the labor baseline, and later AI replica and likeness concerns kept the rights issue visible.

A24 also has a young audience to consider. Roughly 85% of Backrooms’ opening-weekend audience was under 35, and skepticism among younger adults makes AI’s social impact part of the brand risk. Parsons’ criticism of generative AI as a symptom of broader cultural and economic rot sharpens that pressure.

Bringing Parsons and other A24 filmmakers into the process could strengthen the partnership’s credibility or expose where creator-guided language falls short. A24’s brand depends on directors and younger moviegoers who often treat creative identity as part of the product.

A tool that saves planning time still has to look like support for artists, not a shortcut around them. Storyboard images may be a useful early use case because they sit before production, where experimentation is common, but even that use case will be judged by who controls the output and who benefits from the efficiency.

What Would Prove the Partnership Works

Google DeepMind and A24 need usable storyboard applications, not just a demo reel, to show what filmmakers can do differently. Artist control should be visible, and A24’s existing catalog should stay outside training leverage.

For Google’s A24 investment, the next concrete signal is A24 Labs’ first storyboard application in the 2026 partnership: artists need to control the planning images while the no-library-access limit remains intact.