ChatGPT to Limit Emotional Advice Amid Concern of User Dependency

Starting this week, ChatGPT will prompt users to take breaks and reduce direct emotional advice, aiming to prevent unhealthy reliance on the chatbot. OpenAI acknowledged flaws in detecting delusional thinking and is working with doctors and researchers to improve the tool’s responses to sensitive situations. 

These updates follow previous backlash over GPT-4o’s overly agreeable behavior. CEO Sam Altman raised legal concerns over AI-user confidentiality. OpenAI emphasized that success means users achieve their goals-not just spend more time chatting.

“The company wrote in its blog post that it worked with more than 90 physicians across dozens of countries to craft custom rubrics for “evaluating complex, multi-turn conversations.”