AI chatbots have failed people in crisis. Can that be fixed?
AI chatbots are causing documented harm to vulnerable users, particularly those in mental health crises, creating significant legal and reputational liability for technology companies. While AI safety has incrementally improved, critical gaps remain in crisis detection, professional care handoff, and appropriate boundary-setting—requiring greater transparency, third-party evaluation, and clinician involvement in model development. IT leaders must recognize that deploying AI systems without mental health safeguards and explainability creates enterprise risk and erodes public trust.