Senate panel probes Meta AI after Reuters leak on minor policy

A Republican-led Senate panel is preparing to scrutinize Meta Platforms’ generative artificial intelligence after a Reuters disclosure revealed internal guidelines that appeared to permit chatbots to engage in romantic or sensual conversations with minors. The move comes as Meta acknowledged the authenticity of the document but said certain passages were removed after questions from the press, and as lawmakers pressed for a fuller account of what was allowed, when, and by whom.

Sen. Josh Hawley, who chairs the Senate Judiciary Subcommittee on Crime and Counterterrorism, announced an inquiry into whether Meta’s AI products exploited, deceived, or harmed children and whether the company misled the public or regulators about safeguards. In a letter to Meta’s chief executive, Hawley said the panel would demand a complete set of materials related to internal policies, communications, risk reviews, incident reports, and the decision-making trail behind the standards now under review. He urged the company to preserve records and provide them by a September deadline.

The investigation follows a Reuters report detailing an internal policy document, known as the GenAI Content Risk...