A safety partnership, not a treatment approval
On August 6, OpenAI and the American Psychological Association announced plans to develop resources for responsible AI use by young people and families. The work includes practical family guidance, materials for clinicians and school psychologists, and listening to youth, parents and clinicians. OpenAI says more than 260 mental-health experts have contributed to its safety work.
The announcement is not a clinical trial result, a treatment approval or an APA endorsement of ChatGPT as therapy. Effectiveness measures, release dates and independent evaluation methods for the planned resources have not yet been published.
Read it alongside APA’s separate advisory
APA’s official health advisory says general-purpose generative-AI chatbots were not designed as mental-health treatments and should not replace care from qualified professionals. It describes limited evidence for chatbot and wellness-app effectiveness and highlights risks including poor crisis responses, overreliance and privacy exposure.
Practical safeguards
Youth-facing services need age-appropriate explanations, human escalation for crisis signals, data minimization, clear guardian and school roles, and appeal channels. Families should not treat a chatbot response as diagnosis or prescription. Threats of self-harm, violence, abuse or immediate danger require local emergency and qualified professional support.