onversations. OpenAI noted, “ChatGPT may correctly point to a suicide hotline when someone first mentions intent, but after many messages over a long period of time, it might eventually offer an answer that goes against our safeguards.” The company acknowledged that longer conversations can weaken protective measures and said improvements are underway. OpenAI also noted that a portion of conversations will inevitably involve emotional distress. In a related legal development, OpenAI faces a lawsuit from the parents of 16-year-old Adam Raine, who allege that ChatGPT contributed to his death by “actively helping him explore suicide methods.” OpenAI expressed its condolences, saying, “Our deepest sympathies are with the Raine family for their unthinkable loss. Teen wellbeing is a top priority for us.”