# OpenAI Reduces Unsafe ChatGPT Responses by 80% in Mental Health Conversations

OpenAI announced significant improvements to how ChatGPT handles sensitive mental health conversations, working with over 170 mental health experts to make the AI chatbot safer and more supportive.

The collaboration focused on three key areas: helping ChatGPT better recognize when users are in distress, respond with greater empathy, and provide guidance toward real-world professional support resources. According to OpenAI, these changes have reduced unsafe responses by up to 80%.

The update addresses growing concerns about AI chatbots interacting with vulnerable users. As millions of people worldwide use ChatGPT for various conversations—including personal struggles—the potential for harmful or inadequate responses in crisis situations has been a significant safety concern.

By incorporating expertise from mental health professionals, OpenAI aims to ensure that ChatGPT can identify warning signs of dist