OpenAI introduced Trusted Contact, an optional safety tool allowing adults to designate a trusted person who may be notified if ChatGPT detects concerning self-harm discussions. This feature aims to connect users with real-world support during crises, complementing existing crisis hotline integrations. Users can add one trusted adult globally (19+ in South Korea) via settings, who must accept an invitation within a week.

Notifications to contacts include limited context about self-harm concerns and privacy-preserving links to expert guidance, avoiding disclosure of chat details.