Anthropic, OpenAI and Google hold private talks on an industry AI standards body

Anthropic, OpenAI and Google have been holding closed-door working-group meetings since July on creating an industry-led body to set testing and auditing standards for frontier AI models, people familiar with the talks have said, with discussions continuing as recently as last week. CNN independently confirmed the conversations, which began before Anthropic chief executive Dario Amodei’s public call on Sept. 12 for companies to coordinate on safety testing and to slow the pace of capability gains.