ChatGPT Introduces 'Trusted Contact' Feature for Mental Health Safety

OpenAI has launched a new optional safety feature called 'Trusted Contact' within ChatGPT, designed to support users engaging in conversations about self-harm or suicide. This tool allows users to designate an adult contact who will receive a notification if automated systems detect serious safety concerns during chats. The notification provides a generalized reason for the alert rather than a detailed transcript, preserving some privacy while ensuring intervention. To activate the feature, users must manually invite a trusted adult who accepts the request within a week. This development follows research indicating that approximately 1.2 million users weekly discuss suicide-related topics with the AI. The feature complements existing safety measures, such as internal reviews and suggestions for crisis helplines. It also emerges amidst ongoing legal challenges, including a lawsuit from a teen's family alleging ChatGPT contributed to a suicide, and a separate copyright infringement suit filed by PCMag's parent company, Ziff Davis. OpenAI maintains it does not provide instructions for self-harm and continues to refine its safety protocols to address mental health risks associated with AI interactions.