OpenAI, Anthropic and Google DeepMind are collaborating on artificial intelligence safety

OpenAI is collaborating with Anthropic and Google DeepMind to improve artificial intelligence safety, Chris Lehane, the ChatGPT maker’s global policy chief, said, according to Bloomberg and Reuters. The initiative aims to create a mechanism for evaluating and overseeing new AI models amid concerns about the risks associated with the technology’s rapid development.

The proposal was launched by Demis Hassabis, director of Google DeepMind, who suggested establishing an entity modeled on the U.S. Financial Industry Regulatory Authority (FINRA). The body would be funded by companies in the sector and have a board made up of independent experts and representatives of the open-source community.

Under the plan, major AI companies would have to submit their new models to safety evaluations before releasing them on the market. The body could also coordinate a slowdown in development if risk assessments failed to keep pace with technological progress.

In the United States, participation in the evaluation process introduced by the Trump administration at the beginning of August is voluntary. Companies argue that rules are necessary, while the U.S. administration fears that restrictions could benefit China. Such a structure would likely require a legal basis to avoid antitrust issues.