Google is assembling a new team to create advanced AI models capable of simulating the physical world, a step aimed at revolutionizing interactive media and real-time simulations.
Tim Brooks, previously co-lead of OpenAI’s Sora video generator and now with Google DeepMind, shared the announcement on X (formerly Twitter). Brooks, who joined DeepMind in October, will lead this initiative under the umbrella of Google’s AI research division.
“DeepMind has bold plans to develop generative models that replicate the real world,” Brooks wrote on Monday. “I’m building a team dedicated to this mission.”
The new team will collaborate with Google’s existing AI projects like Gemini, Veo, and Genie. Gemini is Google’s flagship AI system for tasks such as text analysis and image generation, while Veo focuses on video creation. Genie, on the other hand, specializes in generating 3D environments and simulating games in real time, with the latest version unveiled in December.
Job postings shared by Brooks describe the team’s goals as addressing “critical new challenges” and scaling AI models “to unprecedented computational levels.” These models are expected to pave the way for applications such as visual reasoning, embodied agent planning, and real-time interactive entertainment.
The team also aims to build tools for “real-time interactive generation” and integrate their models with multimodal systems like Gemini. The overarching vision ties into the pursuit of artificial general intelligence (AGI), a concept referring to AI systems capable of performing tasks on par with human intelligence.
The world modeling field is attracting interest from various startups and major tech players, including prominent researchers like Fei-Fei Lee and companies such as Odyssey and Decart. Advocates envision applications in creating interactive media, developing immersive video games, and designing realistic simulation environments for robotics training.
However, the technology raises concerns. A 2024 study commissioned by the Animation Guild warned that AI could disrupt over 100,000 jobs in U.S. film, TV, and animation industries by 2026. Some companies in this emerging sector, like Odyssey, have pledged to collaborate with creative professionals rather than replace them. It remains to be seen if Google will adopt a similar approach.
Another contentious issue is copyright. Reports suggest some AI models are trained on unlicensed video game footage, potentially exposing developers to legal challenges. While Google asserts that it adheres to YouTube’s terms of service for sourcing training data, the specifics remain undisclosed.
