AI Chatbots Still Roleplay Self-Harm Despite Mental Health Progress
Leading artificial intelligence models from companies like OpenAI, Google, and Anthropic have become significantly less likely to encourage suicide or self-harm in user…