Competency of Large Language Models in Evaluating Appropriate Responses to Suicidal Ideation: Comparative Study

Current versions of 3 major LLMs demonstrated an upward bias in their evaluations of appropriate responses to suicidal ideation; however, 2 of the 3 models performed equivalent to or exceeded the performance of mental health professionals.