OpenAI adds optional Trusted Contact tool to ChatGPT for suicide risk alerts
OpenAI’s ChatGPT now offers an opt-in Trusted Contact feature that can notify a pre-selected person if the bot detects serious suicide risk, but it only works when users set it up in advance.
OpenAI introduced an optional Trusted Contact feature for ChatGPT in May, allowing users to designate a person who will be notified if the system detects severe suicidal ideation. The tool is not active by default; users must enable it in the settings, and it currently does not work with the voice version of the chatbot. When a high-risk cue is identified, the bot may advise the user to set up the contact and will otherwise suggest reaching out to loved ones or calling the 988 crisis line.
Experts such as Holly Wilcox and Ursula Whiteside see value in the added safety net but caution that many at risk may avoid engaging with a bot. Psychologist Jennifer Lockman emphasizes that human connection remains the strongest protective factor and calls for research on the feature’s effectiveness. OpenAI says it is iterating the prompt timing in partnership with mental-health researchers and maintains other safeguards like periodic break reminders and crisis-line referrals.
Why it matters
The feature could provide a timely safety net for users experiencing suicidal thoughts, linking AI detection with real-world help.
In this story
