Chatbots got safer but will still role-play self-harm with users - The Washington Post
September 1, 2026
ChatGPT and other artificial intelligence chatbots have become less likely to encourage suicidal thoughts but will still engage in potentially harmful conversations and reinforce apparently delusional behavior, a new study found.
ChatGPT and other artificial intelligence chatbots have become less likely to encourage suicidal thoughts but will still engage in potentially harmful conversations and reinforce apparently delusional behavior, a new study found.