ChatGPT: 'Trusted Contact' for Self-Harm Safety Alerts

ChatGPT has launched a new "Trusted Contact" feature to help users at risk of self-harm. Here's the thing: adult users can now nominate a friend or family member to be notified if the AI detects serious self-harm risk during a conversation. This optional setting is rolling out globally for users aged 18 and up, and 19 in South Korea. What's interesting is how it works. If ChatGPT's automated systems flag concerning content, it will first prompt the user to reach out to their chosen contact. Then, trained human reviewers assess the situation. If they confirm a serious safety concern, the Trusted Contact might get an email, text, or in-app alert. This alert is limited; it won't share chat logs but encourages the contact to check in. Every alert gets human review, ideally within an hour. This feature adds another layer of support, connecting users with real-world help when they need it most.