WorkThe story, in brief
Strengthening ChatGPT’s responses in sensitive conversations
80% reduction in unsafe responses. OpenAI just rebuilt ChatGPT's mental health guardrails with 170+ experts.
The KeyNews take
Why it matters
AI safety in high-stakes domains is moving from theoretical ethics debates to measurable deployment. This signals how serious AI labs are getting about responsible release in sensitive use cases—and sets a new standard for competitor safety claims.
The key facts
5 to knowOpenAI collaborated with 170+ mental health experts
Unsafe responses reduced by up to 80%
Focus on distress recognition, empathetic response, real-world support guidance
Published Oct 27, 2025
Safety governance in mental health/crisis support domain
Go to the source
OpenAI Blogopenai.com
Publisher excerpt: OpenAI collaborated with 170+ mental health experts to improve ChatGPT’s ability to recognize distress, respond empathetically, and guide users toward real-world support—reducing unsafe responses by up to 80%. Learn how we’re making ChatGPT safer and more supportive in sensitive moments.