OpenAI introduced an optional feature called Trusted Contact in May, designed to connect ChatGPT users experiencing suicidal thoughts or self-harm concerns with trusted individuals during emotional crises. According to NPR estimates based on data released by OpenAI, more than a million weekly users discuss suicide and self-harm on the platform. The feature functions by having ChatGPT detect warning signs of serious safety risks and alerting a user’s designated contact if such indicators emerge.
Despite the feature’s existence, many users remain unaware of it. When NPR tested how ChatGPT communicates the feature to new users, a first-time user received no initial notification about Trusted Contact upon signing up. Only when directly asked did the chatbot confirm the feature exists and explain that it must be manually activated through the app’s settings. OpenAI indicated the company has observed “good adoption” but did not release specific usage numbers. A company spokesperson noted that the feature currently functions through text-based chat but is not yet available on ChatGPT’s voice interface, though the company is working toward that capability.
Suicide prevention experts view the feature with cautious optimism. Holly Wilcox, director of the Center for Suicide Prevention at Johns Hopkins Bloomberg School of Public Health, suggested it would be beneficial if offered to all users proactively before emergencies arise. Ursula Whiteside, CEO of suicide prevention nonprofit Now Matters Now, tested the setup process and found it relatively straightforward, though she acknowledged it might prove challenging for someone in acute distress. Focus groups conducted by Whiteside’s organization indicated that individuals with lived experience of suicidal thoughts valued the potential for establishing trusted connections in advance during clearer mental states.
However, a critical limitation exists: the feature only functions if users set it up beforehand. If someone in crisis has not previously configured their Trusted Contact, ChatGPT cannot notify anyone on their behalf. In such situations, the chatbot can only provide supportive responses encouraging the user to contact loved ones or the National Suicide and Crisis Lifeline at 9-8-8.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI