ChatGPT Adds Trusted Contact Feature for Mental Health

OpenAI is introducing a new safety feature that could help protect ChatGPT users during vulnerable moments.

The company announced last week that it's developing a "trusted contact" option that allows adults to designate someone who'll receive notifications if the chatbot detects signs of emotional distress. Think of it like an emergency contact, but for mental health.

The feature comes after extensive work with OpenAI's Council on Well-Being and AI, a group of experts assembled to address mental health concerns related to chatbot use. While details remain limited, the system would monitor conversations for potential crisis indicators and alert the user's chosen contact when additional support might be needed.

OpenAI says millions of its weekly users show signs of suicidality, psychosis, or other mental health concerns during their conversations. The company is also developing new evaluation methods that simulate extended mental health discussions to better identify risks and improve how ChatGPT responds in sensitive situations.

The tool could be especially valuable for people managing diagnosed mental illnesses who understand that intensive AI use might intersect with their mental health in challenging ways. Some users have successfully managed conditions for years before experiencing difficulties related to chatbot interactions.

The Bright Side

This feature represents a thoughtful approach to user safety in an emerging technology space. By giving people the option to create a safety net, OpenAI acknowledges both the reality of how people use chatbots and the importance of human connection during difficult times.

The notification system respects user autonomy while offering protection. People can choose whether to activate it, who to designate, and maintain privacy unless the system detects genuine concern.

OpenAI's announcement also highlighted that it's "continuing to advance how our models detect and respond to signs of emotional distress." The company is working on better identifying potential risks and improving ChatGPT's responses during sensitive conversations.

While questions remain about how the system will define crisis moments and what happens if users don't opt in, the trusted contact feature shows tech companies taking concrete steps toward responsible AI development. It puts safety tools directly in users' hands.

The feature should become available as OpenAI continues developing and refining the notification standards.