Testing Reveals ChatGPT for Teens Safety Gaps During Crises
October 7, 2026
New evaluations show that ChatGPT's teen-specific safeguards may fail to prevent engagement during mental health crises. Findings suggest the model may encourage unhealthy user-AI relationships despite existing protections.
HOW THIS AFFECTS YOU
●
policyThis signals a need for more rigorous safety evaluations regarding emotional dependency in minors.