Use of Generative AI Chatbots and Wellness Applications for Mental Health

Key Findings

• — The ability of general-purpose generative-AI chatbots to consistently and safely manage a user in crisis is characterised as limited and unpredictable.
• — Such tools can foster unhealthy dependence by blurring the line between interaction with a digital tool and human connection, and can act as amplifiers of pre-existing vulnerabilities.
• — The advisory states these tools should not replace qualified mental-health providers, and calls for clinical trials, independent evaluation, privacy safeguards, and clinician AI literacy.

Methodology Notes

Institutional health advisory / expert-panel product; individual authors not named on the landing page (authors left empty, gap noted). Landing page dated November 2025; the day (13 November 2025) is from the accompanying APA press release. A full-text PDF is linked from the advisory landing page. Distinct from the APA 'Artificial Intelligence and Adolescent Well-Being' advisory (June 2025) already held — this advisory addresses general-purpose GenAI chatbots and wellness apps for mental health across all ages, not adolescents specifically.

Sources