- The case, lodged in a California court last week, alleges that the chatbot encouraged the teenager’s suicidal thoughts and even helped him prepare notes before taking his life.
- OpenAI has acknowledged that its current system sometimes falters in lengthy interactions and says it is working to ensure consistent safety protocols throughout extended sessions.
- OpenAI said the changes will be incorporated into its next major model update, GPT-5, which is expected to power the chatbot later this year.