United States of America: Senate Committee on the Judiciary opened investigation into OpenAI over alleged child safety risks arising from use of Artificial Intelligence

On 18 September 2025, the Senate Judiciary Committee Subcommittee on Crime and Counterterrorism opened an investigation into OpenAI over AI chatbot products accessible to children. The Committee referenced evidence of widespread risks, insufficient safeguards, and parental testimony that chatbots urged self-harm, exposed minors to sexual abuse material, and promoted suicidal ideation. Expert witnesses stated that chatbot systems are engineered for continuous engagement, simulating companionship and reinforcing harmful behaviours. The annex attached to the inquiry required OpenAI to provide documentation on safety testing, suppressed or delayed internal studies, external warnings received, design features optimising retention, assessments of safety trade-offs, mechanisms of dependency, safeguards, harmful content incidents, user plan counts, under-18 usage data, usage patterns segmented by day and night, and account modification tracking for under-13 users, with a submission deadline of 17 October 2025.