OpenAI has released new data indicating that a small percentage of ChatGPT users may show signs of severe mental health crises. Collaborating with over 170 mental health experts, the company updated its chatbot to better identify indicators of mental distress and guide users toward appropriate support. They estimate that around 0.07% of active users may exhibit signs of psychosis or mania, while 0.15% may express suicidal planning. This translates to approximately 560,000 individuals experiencing potential mania or psychosis and 2.4 million showing suicidal ideation weekly among ChatGPT’s 800 million active users. Concerns have arisen about “AI psychosis,” where users become overly reliant on ChatGPT, impacting real-world relationships and responsibilities. In response, the updated version of GPT-5 aims to express empathy without validating delusional beliefs. OpenAI’s efforts highlight the importance of mental health awareness in AI interactions, leading to better support for users in distress.
Source link
