OpenAI wants users to be less emotionally dependent on ChatGPT with new mental well-being features

ChatGPT

OpenAI is introducing “gentle reminders” in ChatGPT to nudge users about how long they’ve been using the service, encouraging them to take breaks and return to their lives after completing their tasks. The move is part of OpenAI’s broader philosophy that success isn’t measured by time spent on ChatGPT, but by helping users accomplish their goals efficiently.

“We built ChatGPT to help you thrive — to make progress, learn something new, solve a problem — and then get back to your life,” OpenAI said in a statement. “Our goal isn’t to hold your attention, but to help you use it well.”

Beyond managing screen time, OpenAI is also working to refine ChatGPT’s responses to sensitive, high-stakes questions. Instead of offering direct advice on personal dilemmas like “Should I break up with my boyfriend?”, ChatGPT will shift to helping users think through the pros and cons, encouraging more thoughtful decision-making.

Furthermore, OpenAI is enhancing ChatGPT’s ability to recognise signs of mental or emotional distress. The AI will avoid amplifying harmful delusions or emotional dependencies and instead direct users to evidence-based resources.

To bolster these initiatives, OpenAI has collaborated with over 90 physicians across 30 countries, developing detailed rubrics for complex conversations. Human-computer interaction researchers are also stress-testing ChatGPT’s responses, ensuring the model can identify concerning behaviours. Furthermore, an advisory group of experts in mental health, youth development, and AI-human interaction is being formed to guide future safety measures.