Google DeepMind's New AI World Models Team, Led by Tim Brooks
- Google DeepMind is building AI world models for real-time simulation, led by Tim Brooks and collaborating with Gemini, Veo, and Genie.
- Concerns about AI adoption include job disruption in creative industries and copyright issues over unlicensed video game playthroughs used in training data.
Google is advancing its artificial intelligence capabilities with the formation of a new team dedicated to building AI models that simulate the physical world. Announced by Tim Brooks, a former co-lead of OpenAI’s video generator Sora and now a leader at Google DeepMind, the initiative will focus on creating generative models for real-time, interactive simulations.
“DeepMind has ambitious plans to make massive generative models that simulate the world,” Brooks wrote on X, adding, “I’m hiring for a new team with this mission.”
Brooks’ team will collaborate with existing Google AI teams, including those behind Gemini, Veo, and Genie, according to job listings linked in his post. Gemini, Google’s flagship AI model series, handles tasks like text generation and image analysis. Veo specializes in video generation, while Genie focuses on real-time 3D environment simulation. Google’s latest Genie model, unveiled in December, can generate expansive and playable 3D worlds. Google DeepMind revealed Genie 3 in August 2025, describing it as a real-time interactive general-purpose world model. Veo 3.1 remains the current Veo model and includes 4K upscaling, native 9:16 vertical video, and advanced video-generation features.
The new team aims to address “critical new problems,” scale AI models to higher computational capacities, and develop tools for “real-time interactive generation.” This aligns with broader aspirations for artificial general intelligence (AGI) — AI capable of performing tasks on par with human abilities. The models will support various applications, including visual reasoning, robotic planning, and interactive entertainment.
Newsletter
The corridor, every morning.
Funding rounds and cross-border capital moves, one email, five minutes.
While other firms like World Labs, Decart, and Odyssey are also exploring world models for media and simulation purposes, concerns remain. Creatives worry about potential job disruptions, as a Wired investigation highlighted AI adoption in game studios to replace workers. Additionally, copyright challenges loom, with questions surrounding the use of unlicensed game playthroughs for model training.
Google, leveraging YouTube’s terms of service, asserts permission to train its models on platform videos but hasn’t disclosed specific sources. It remains to be seen whether the company will address these concerns or adopt a collaborative approach with creatives, as some startups have pledged.
More in News
