AI RESEARCH
AI Safety Training Can be Clinically Harmful
arXiv CS.LG
•
ArXi:2604.23445v1 Announce Type: cross Large language models are being deployed as mental health agents at scale, yet only 16% of LLM-based chatbot interventions have undergone rigorous clinical efficacy testing, and simulations reveal psychological deterioration in over one-third of cases. We evaluate four generative models on 250 Prolonged Exposure (PE) therapy scenarios and 146 CBT cognitive restructuring exercises (plus 29 severity-escalated variants), scored by a three-judge LLM panel.