OpenAI Boosts ChatGPT Safety After 1M Weekly Suicide Queries

OpenAI reports that ChatGPT processes over 1.9 million self-harm queries weekly, including more than 1 million suicide-related messages. ChatGPT sessions have a median length of 33 messages. To improve AI safety and mental health support, the company updated its GPT-5 model after consulting over 170 experts. The update boosts appropriate responses by 65% and raises compliance with suicide prevention guidelines to 91%. It also adds clarifying questions, empathetic replies, and crisis hotline contacts. Enhanced moderation filters detect high-risk language, escalate cases for human review, and introduce benchmarks for emotional reliance, non-suicidal crises, parental controls and age detection. Amid a lawsuit and state AG warnings, OpenAI’s responsible AI measures underline its commitment to user safety. Traders should note the growing role of AI in compliance and tech regulation, while the broader market impact remains neutral.