Google is forming a new team to work on AI models capable of simulating the physical world, marking a significant leap in artificial intelligence research.
Tim Brooks, a former co-lead on OpenAI’s video generator, Sora, who joined Google DeepMind in October, announced his leadership of a new team in a post on X. initiative will be housed under Google DeepMind, the company’s AI research lab.
“DeepMind has ambitious plans to make massive generative models that simulate the world,” Brooks stated. “I’m hiring for a new team with this mission.”
The new modeling team will build upon existing work by Google’s Gemini, Veo, and Genie teams. According to job listings linked in Brooks’ announcement, the team aims to address “critical new problems” and scale AI models “to the highest levels of computing.”
Gemini represents Google’s flagship AI models for tasks such as image analysis and text generation. Veo focuses on video generation, while Genie specializes in creating AI-driven simulations of games and 3D environments. Genie’s latest model, previewed in December, showcased its ability to generate a vast variety of playable 3D worlds.
Advancing Toward Artificial General Intelligence (AGI)
Google DeepMind sees scaling AI training on video and multimodal data as crucial to achieving artificial general intelligence (AGI), defined as AI capable of performing tasks at the level of a human. According to job descriptions, the new team will explore applications of world models across domains such as visual reasoning, simulation, planning for embodied agents, and real-time interactive entertainment.
The team’s work will also focus on “real-time interactive generation” tools and integrating IR models with multimodal systems like Gemini.
Competitive Landscape and Challenges
Several startups and major tech firms, including Fei-Fei Lee’s World Labs, Israeli firm Decart, and Odyssey, are also pursuing world modeling technologies. Se models hold promise for applications ranging from video games and movies to realistic training simulations for robots.
However, the technology faces criticism and concerns. A Wired investigation revealed that companies like Activision Blizzard are using AI to streamline production, sometimes at the expense of jobs. A 2024 study by Animation Guild projected that over 100,000 U.S. film, television, and animation jobs could be disrupted by AI by 2026.
Some startups, like Odyssey, have pledged to collaborate with creative professionals rather than replace m. Where Google will adopt a similar approach remains to be seen.
Ethical and Legal Questions
rise of world models also brings unresolved copyright issues. Some models are trained on video game playthroughs and or content, raising concerns about the legality of using unlicensed material. Google, which owns YouTube, asserts its right to train models on videos by YouTube’s terms of service but has not disclosed specific sources for training data.
Google DeepMind’s initiative underscores growing interest and investment in AI technologies that can simulate the physical world. As the field advances, balancing innovation with ethical and legal considerations will remain a pivotal challenge.
