Major AI companies have been in discussions about forming their own standards body. Anthropic, Google, and OpenAI are among the companies involved in these conversations. The talks were underway before a recent resignation by a former Anthropic researcher, who criticized the companies for not behaving responsibly.

The idea of a standards body was also proposed by Google DeepMind founder Demis Hassabis in a July essay. He suggested a US-led body that would test advanced AI models before they are deployed.

Further details about the proposed standards body and the ongoing discussions are available from the source, which initially reported on the conversations through The Information and people familiar with the situation.