The new team will be led by Tim Brooks, who revealed it on X. Brooks was a co-lead on OpenAI’s video generator, Sora, and departed for Google’s AI research centre, Google DeepMind, in October. It will be a component of DeepMind on Google.
DeepMind has ambitious plans to make massive generative models that simulate the world. I'm hiring for a new team with this mission. Come build with us!https://t.co/pqvALtAvLshttps://t.co/vtwgeXl9Dl
Under Tim Brooks’ leadership, Google is assembling a new artificial intelligence team within its DeepMind division to develop models that can replicate the real world. With the goal of creating real-time interactive generation tools, this team will build upon the work of Google’s current AI models, including Gemini, Veo, and Genie. The idea and initiative draws attention to the possible advantages and moral dilemmas of using AI to simulations and the creative industries.
Brooks said on Monday morning, “DeepMind has ambitious plans to make massive generative models that simulate the world.” “I’m hiring for this mission on a new team.”
The new modelling team will work with and build on work from Google’s Gemini, Veo, and Genie teams to address “critical new problems” and scale models “to the highest levels of compute,” per jobadvertisements that Brooks cited in his post. Veo is Google’s proprietary video generation model, whereas Gemini is the company’s main line of AI models for tasks including text generation and picture analysis.
Genie is Google’s version of a world model, which is an AI capable of real-time game and 3D environment simulation. A vast array of playable 3D worlds may be created using Google’s most recent Genie model, which was shown off in December.
One of the job descriptions states, “We think scaling [AI training] on video and multimodal data is on the critical path to artificial general intelligence.” In general, artificial general intelligence, or AGI, refers to AI that is capable of performing any task that a person can. “Many domains, including visual reasoning and simulation, embodied agent planning, and real-time interactive entertainment, will be powered by world models.”
According to the description, Brooks’ new team will investigate how to link their models with pre-existing multimodal models like Gemini and seek to create “real-time interactive generation” tools on top of the models they construct.
Influential AI researcher Fei-Fei Lee’s World Labs, Israeli upstart Decart, and Odyssey are among the many startups and large tech businesses vying for world models. They think that world models may eventually be utilized to produce realistic simulations, such as robot training settings, and interactive media, such as video games and movies.
Creatives, however, feel differently about the technology.
Game companies like Activision Blizzard, which has fired dozens of employees, are utilizing AI to reduce costs, boost output, and make up for attrition, according to a recent Wired study. Additionally, a 2024 research commissioned by the Animation Guild, an organization that represents cartoonists and animators in Hollywood, predicted that by 2026, artificial intelligence would threaten more than 100,000 jobs in the film, television, and animation industries in the United States.
Odyssey and other firms in the emerging area of modelling have promised to work alongside creative professionals rather than take their place. We’ll see whether Google does the same.
Additionally, the issue of copyright remains unanswered. The fact that certain world models seem to have been taught using video game playthrough tapes could lead to legal action against the corporations that created them in situations where the films were not licensed.
YouTube’s parent company, Google, claims that it is authorized to use YouTube videos to train its models in compliance with the platform’s terms of service. However, the business has not specified which particular videos it is using for training.
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behaviour or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
