OpenAI is adding a safety net to ChatGPT that lets adult users designate a "Trusted Contact" - a friend, family member, or caregiver who gets notified if the AI detects concerning conversations about self-harm or suicide.
The feature is entirely opt-in. Users choose who to designate, and that contact only receives an alert if ChatGPT's systems flag a conversation as potentially dangerous. OpenAI hasn't published specifics about exactly what triggers a notification, but the framing points to explicit discussions of suicide or self-harm - not general expressions of stress or a rough week.
What OpenAI Is Actually Acknowledging
This is an implicit admission that a significant number of people are using ChatGPT for emotional support - and that some of those conversations get serious. That's not new information; therapists and researchers have raised concerns about AI as an emotional outlet for two years. But OpenAI building an active safety mechanism into the product is a meaningful shift from "we're a general-purpose assistant" toward acknowledging what people actually use the thing for.
The privacy tradeoff here is real. By enabling this feature, users are telling OpenAI: monitor my conversations for danger signals and report them to a third party. That's a significant grant of access. The opt-in design handles the consent question reasonably - no one gets enrolled without choosing to. But it also means the people most at risk may be least likely to set it up in the first place.
A Safety Layer, Not a Therapy Replacement
Crisis lines, hotlines, and mental health apps have all grappled with the same tension: how do you respond when someone is clearly struggling without overreaching? OpenAI's approach borrows from real-world crisis intervention - loop in a trusted person who can provide actual human follow-up.
A designated contact receiving a notification is better than nothing, but it isn't a substitute for professional support. It won't connect someone to a therapist or ensure they get care. What it does is reduce the chance that a dangerous conversation happens in complete isolation.
The Trusted Contact feature is rolling out to adult users. OpenAI has not announced a timeline for broader availability or said whether it will eventually expand to minors.