OpenAI is adding a new “Trusted Contact” feature to ChatGPT after mounting scrutiny over how AI chatbots handle conversations involving self-harm and suicide.
The company announced Thursday that adult users can now assign a trusted person, like a parent, sibling, partner, or friend, to their account. If ChatGPT detects signs of serious self-harm risk during a conversation, the system may encourage the user to contact that person directly while also sending the trusted contact a brief alert.
OpenAI says the notification will avoid sharing private chat details and will only encourage the contact to “check in” with the user. The company also said potentially dangerous conversations are reviewed by both automated systems and human safety teams. “We strive to review these safety notifications in under one hour,” OpenAI stated.
The rollout arrives as AI companies face growing legal pressure tied to mental health harms involving chatbot interactions. Several lawsuits filed against OpenAI and Character.AI allege that vulnerable users, including teenagers, were exposed to harmful conversations involving suicide, emotional dependency, or self-harm discussions.
One of the most widely reported cases involved 14-year-old Sewell Setzer III, whose family alleged a Character.AI chatbot encouraged emotional attachment before his death by suicide. Another lawsuit filed against OpenAI claims ChatGPT allegedly validated suicidal thoughts expressed by 16-year-old Adam Raine before his death in 2025.
Researchers have also warned about the psychological risks tied to AI companionship platforms. Multiple academic studies published over the last two years found that emotionally dependent chatbot use may worsen isolation, encourage unhealthy attachment, and increase risks for vulnerable users already dealing with mental health struggles.
“Trusted Contact is part of OpenAI’s broader effort to build AI systems that help people during difficult moments,” the company wrote in its announcement.
