FDA has not authorized AI chatbots for mental health care
2 min readNearly one-fifth of Americans aged 12 to 21 reported using an AI chatbot for mental health advice in 2025. The Government Accountability Office said in a report published Friday 20 November 2026 that experts warn general-purpose chatbots are not designed for such needs and can be dangerous for this purpose.
As of November 2026, the Food and Drug Administration had not granted marketing authorization for any chatbots for mental health care. The FDA has not authorized any AI chatbot to support mental health. Purpose-built chatbots are also not a replacement for qualified mental health providers.
General-purpose chatbots allow users to ask about a wide range of topics. They tend to be agreeable and flattering and may reinforce a user's perspective rather than challenge it. Purpose-built chatbots are trained on mental health data and therapeutic techniques, according to developers. Both types can generate confident-sounding responses that may not be accurate or understand personal context.
The agency said chatbots do not have the clinical judgment of a mental health professional. General-purpose chatbots have been linked to individuals harming themselves, leading some developers to implement guardrails. Experts warned that some guardrails may be ineffective and may not keep pace with technological advances. Chatbot regulation is also evolving at the state and federal level.
Some developers claim their purpose-built chatbots use therapeutic approaches. Experts cautioned that such claims may lead users to believe the chatbots provide care comparable to a mental health professional. Some companies reported that their chatbots have undergone or are undergoing randomized controlled trials. Experts said more high-quality studies are needed and it is too early to substitute chatbots for professional care.
Purpose-built chatbots could help address access gaps caused by high costs of care, provider shortages and stigma. Commonly used general-purpose chatbots may identify mental health resources and connect users with qualified providers. Risks include harm to at-risk populations, reinforcing delusions or negative feelings, uncertain effectiveness, inadequate regulation and privacy risks from retained conversation data. The agency called for research on safety, standards, guardrails, monitoring and responsibility for negative outcomes.



