OpenAI announced the addition of an optional safety feature, 'Trusted Contact,' to ChatGPT on May 7, 2026. This feature allows adult users to register a trusted individual, such as a friend or family member, who will be notified if the system detects conversations indicating a serious risk of self-harm.

When the automated monitoring system identifies a risk, ChatGPT first warns the user of a potential notification and provides suggested conversation starters. Subsequently, trained human reviewers evaluate the situation and, if necessary, contact the designated person via email or text message. OpenAI aims to complete these reviews within one hour.

Alerts sent to contacts disclose only that a safety concern exists, omitting chat transcripts or specific details to maintain user privacy. Developed with guidance from clinicians, suicide prevention experts, and the American Psychological Association (APA), the feature serves as an additional measure to facilitate real-world emergency support.

OpenAI announced the addition of an optional safety feature, 'Trusted Contact,' to ChatGPT on May 7, 2026. This feature allows adult users to register a trusted individual, such as a friend or family member, who will be notified if the system detects conversations indicating a serious risk of self-harm.

When the automated monitoring system identifies a risk, ChatGPT first warns the user of a potential notification and provides suggested conversation starters. Subsequently, trained human reviewers evaluate the situation and, if necessary, contact the designated person via email or text message. OpenAI aims to complete these reviews within one hour.

Alerts sent to contacts disclose only that a safety concern exists, omitting chat transcripts or specific details to maintain user privacy. Developed with guidance from clinicians, suicide prevention experts, and the American Psychological Association (APA), the feature serves as an additional measure to facilitate real-world emergency support.