OpenAI Discloses Staggering Mental Health Metrics Among ChatGPT Users

OpenAI's latest transparency report reveals alarming mental health trends among ChatGPT users. Approximately 1.2 million weekly users—representing 0.15% of the platform's 800 million user base—engage in conversations about suicide with the AI assistant. Nearly 400,000 exhibit explicit suicidal intent, while another 560,000 show signs of psychosis or mania.

The company acknowledges the challenge in detecting these rare but critical interactions. GPT-5's safety mechanisms now intercept 91% of such cases, marking significant improvement over previous models. However, the sheer volume of distressed users has sparked renewed ethical scrutiny around AI's role in mental health crises.

Separately, Decrypt's SCENE hub continues to expand its coverage of Web3 culture, though no direct cryptocurrency correlation appears in this mental health disclosure. The data underscores the growing complexity of managing large-scale AI platforms where human vulnerability intersects with algorithmic responses.