ChatGPT: New self-harm alert feature for trusted contacts

OpenAI's ChatGPT can now alert a trusted friend or family member if it detects a user is at serious risk of self-harm. This new "Trusted Contact" feature adds a layer of human intervention to the AI's safety measures. Here's the thing: adult users can nominate a friend or family member through ChatGPT’s settings. If the AI flags a conversation about potential self-harm, trained human specialists will review it. What's interesting is that notifications are not automatic; human teams assess the risk before any alert goes out via email, text, or in-app message. The designated contact must be over 18 and accept the role. This builds on existing safety tools and aims to provide an additional level of support beyond crisis helplines. The bottom line: this feature highlights the growing debate about AI's responsibility when users discuss sensitive topics.