OpenAI is updating ChatGPT's mental health safeguards, and its legal exposure is part of the story.
The company outlined new and revised features aimed at protecting users who may be struggling. Parental controls give families more visibility into how minors interact with the platform. A trusted-contacts feature lets users designate someone to notify if they appear to be in crisis. OpenAI also says it has improved the model's distress detection, though the company did not publish benchmarks or error rates for that system. The update also acknowledged recent litigation: lawsuits alleging ChatGPT contributed to harm for vulnerable users.
Mental health has become one of the most contested fronts in the AI industry, following several high-profile cases linking chatbot interactions to self-harm. The problem with distress detection is that "improved" is an unverifiable claim without independent testing. A model that gets better at sounding empathetic is not the same as one that reliably catches a user in crisis.
For a company expanding ChatGPT into healthcare and education, adding safety optics while facing lawsuits is a reasonable move, just not necessarily a reassuring one.