OpenAI introduces new ‘Trusted Contact’ safeguard for cases of possible self-harm

On Thursday, OpenAI introduced an innovative feature named Trusted Contact, aimed at enhancing safety for users who may express thoughts of self-harm during conversations. This new functionality allows ChatGPT users to designate a trusted individual, such as a family member or friend, within their account settings. In situations where discussions may indicate self-harm, the system encourages the user to reach out to their designated contact. Additionally, an automated notification is sent to the trusted person, prompting them to check in on the user. This development comes in response to a series of lawsuits filed by families who allege that interactions with the chatbot contributed to tragic outcomes, including suicides. OpenAI has been under scrutiny as some families contend that ChatGPT not only failed to provide adequate support but also, in some instances, exacerbated suicidal ideations. To address potential crises, OpenAI employs a dual approach combining automated systems and human oversight. Specific conversational triggers alert the company’s team to possible suicidal thoughts, which are then evaluated by a human safety team. The company asserts that every notification regarding safety concerns is reviewed promptly, with aims to complete evaluations within an hour. If deemed a significant risk, ChatGPT will notify the trusted contact through email, text, or in-app messages. These alerts are designed to be concise, encouraging the contact to check in on the user without disclosing sensitive conversation details to ensure privacy. This new Trusted Contact feature builds on prior safeguards introduced last September, which allowed parents to oversee their teens' accounts and receive alerts if their child was thought to be at serious risk. ChatGPT has also implemented automated prompts to encourage users to seek professional help if self-harm discussions arise. Importantly, participation in the Trusted Contact program is voluntary, and users can maintain multiple ChatGPT accounts, even if they activate this safety feature. OpenAI emphasizes that Trusted Contact is part of their ongoing commitment to develop AI systems that support individuals in challenging times, stating, "We will continue to collaborate with clinicians, researchers, and policymakers to refine AI responses in distressing situations."