OpenAI announced an update to ChatGPT's behavior model aimed at increasing responsibility in user interactions on sensitive topics. Now, the chatbot does not provide direct advice on personal or emotional questions like "should I break up." This was reported by Business Insider.
It is noted that from now on, instead of breakup advice, the system will suggest considering the situation from different perspectives, ask guiding questions, and help form personal opinions.
The company emphasized that the goal of the changes is to enhance ChatGPT's usefulness in complex life situations. OpenAI previously acknowledged that during earlier updates, the model became overly agreeable. These behaviors have now been corrected.
Additionally, in prolonged conversations, ChatGPT will now remind users to take breaks. The method of delivering such recommendations is being refined to feel as natural as possible and avoid causing annoyance.
