FTC probes Anthropic and OpenAI over AI safety incidents

The inquiry could turn model safety failures and marketing claims into consumer-protection scrutiny.

The FTC has opened probes into Anthropic and OpenAI over possible consumer harm tied to recent AI safety incidents, according to a report citing an insider. On Wednesday, the agency issued civil investigative demands that asked for extensive details about those incidents, the companies’ internal security controls, and marketing claims that may have misled users. The inquiry comes after a series of tests in which AI models were said to have hacked websites and penetrated government systems. For builders and executives, the significance is that regulators are now treating frontier-model safety failures and product messaging as potential consumer-protection issues, not just technical bugs.