Google is forming a new team to build AI that can simulate the physical world
Google DeepMind forms a new team creating AI models simulating the world.

Google DeepMind is establishing a new team, led by Tim Brooks, to develop AI models capable of simulating the physical world. This initiative aims to build on the work from Google's Gemini, Veo, and Genie teams, tackling new challenges and significantly scaling models, as announced in Brooks' post on X.
Their focus includes creating real-time interactive generation tools and integrating these models with existing multimodal models, like Gemini. Many startups and tech companies, including Fei-Fei Lee's World Labs and Odyssey, are exploring world models for applications in interactive media and simulations for training robots.
Concerns arise regarding the potential disruption of creative jobs, with studies predicting substantial impacts by 2026. Additionally, issues surrounding copyright, particularly related to video game clips used for training models, remain unresolved, although Google claims to comply with YouTube’s terms for training data.