Google DeepMind Assembles Team for World-Simulating AI Models

Google is forming a new team within its AI research lab, Google DeepMind, to work on AI models capable of simulating the physical world. Tim Brooks, former co-lead of OpenAI’s video generator Sora, will lead the team.

The Team’s Mission

Brooks shared the news in a post on X, outlining the team’s goal to build massive generative models that simulate the world.

The team will work closely with existing Google AI teams like Gemini, Veo, and Genie to tackle new challenges and scale models to the highest levels of computing.

The Role of World Models

World models are AI systems that simulate real-time 3D environments and games. Google’s latest Genie model can generate a variety of playable 3D worlds, with applications in visual reasoning, planning for embodied agents, and real-time interactive entertainment.

Industry Impact

The AI world modeling field is rapidly growing, with companies like Odyssey and Decart also working on similar projects.

However, creatives have mixed feelings about the technology, as it could disrupt jobs in industries like gaming and animation.

Some startups, including Odyssey, aim to collaborate with creative professionals rather than replace them.

Copyright Concerns

There are ongoing concerns regarding the copyright of video content used to train these models.

Google asserts that it sources videos from YouTube in accordance with the platform’s terms of service, though it has not revealed which specific videos are being used.