Policy

US watchdog says no mental-health AI chatbot has FDA authorization in new assessment

A November 20 GAO technology assessment finds no FDA-authorized mental-health chatbot and highlights uncertainties over effectiveness, safeguards and the use of such tools by young people.

Updated: 2 min read

The US Government Accountability Office reported on November 20, 2026, that the Food and Drug Administration had not granted marketing authorization to any artificial-intelligence chatbot for mental-health care as of November. Its Science and Technology Spotlight examines the growing use of conversational tools for emotional support and the uncertainties surrounding their effectiveness and safety.

The assessment says nearly one-fifth of Americans aged 12 to 21 reported using an AI chatbot for mental-health advice in 2025. That figure describes reported use for advice, not the proportion receiving clinical treatment or the number harmed. GAO distinguishes general-purpose chatbots from purpose-built applications designed for narrower tasks.

The watchdog identifies several areas of concern, including systems that agree with users inappropriately, guardrails that may not consistently work, uncertain clinical effectiveness and evolving regulatory oversight. It also notes potential benefits, such as helping people find resources and complementing professional care where access is limited. GAO does not present chatbots as substitutes for qualified mental-health providers.

The report says some developers have advertised FDA designations that permit accelerated consideration for specific potential medical uses. Such designations do not themselves constitute authorization to market a chatbot for mental-health care. Independent coverage by the legal firm Alston & Bird and CASRAI described the report's findings and its distinction between potential support tools and clinically authorized treatments.

The findings concern technology assessment and regulatory status, not a determination that any named provider committed a violation. Future clinical evidence, product-specific claims and regulatory decisions will determine how individual applications are evaluated. The assessment underscores why availability in an app store should not be confused with demonstrated medical effectiveness or FDA authorization.

Spotted an error? Request a correction · Our editorial principles · Contact the newsroom