Google, OpenAI and Anthropic reportedly discuss industry-run AI safety body

Google, OpenAI and Anthropic have reportedly been discussing an industry-run body for AI testing and auditing, with working-group meetings continuing since July. The Information says a launch could come in late 2026 or 2027 and the companies are considering potential CEOs. The talks follow a stalled White House effort to support a similar organization. OpenAI CEO Sam Altman reportedly told employees that the labs may have to proceed without government backing. No organization has launched, and the report says Meta’s position remains unclear.