1 of 1
Story summary
- OpenAI released data showing ChatGPT users experience mental health issues, with 0.15% showing signs of potential suicidal intent (over a million weekly) and 560,000 showing signs of mania or psychosis.
- OpenAI updated its GPT-5 model to address these concerns, claiming a 91% compliance rate in handling suicidal conversations.
- Despite improvements, OpenAI acknowledges ongoing challenges with mental health issues among users.
- OpenAI is implementing safeguards, including age detection for child users.
