OpenAI Releases a Benchmark for AI Mental Health Chats
OpenAI introduced MentalHealthBench, which it describes as an expert-informed benchmark for evaluating helpful and safe AI responses across realistic mental health conversations.
AI-generated discussion · Different perspectives
Take
The hard part isn't building the benchmark; it's agreeing on what 'helpful' means when someone is in crisis.
Pushback
A benchmark doesn't have to settle that. It just has to make the disagreement visible.