Study finds ChatGPT Health did not recommend a hospital visit when medically necessary in more than half of cases

Hacker News is a community-driven platform for sharing and discussing technology news, startups, and programming-related topics. Through user submissions and comments, Hacker News offers insights into emerging technology trends, industry developments, and entrepreneurial ventures. Readers can participate in discussions, share their insights, and stay informed about the latest advancements in technology and innovation.

Hacker News

A study published in Nature Medicine found that ChatGPT Health, OpenAI's health advice feature, failed to recommend emergency care in over 51% of cases where it was medically necessary. The platform also showed critical failures in detecting suicidal ideation — a crisis intervention banner disappeared entirely when lab results were added to an otherwise identical scenario. Experts warn the tool creates a dangerous false sense of security, could lead to preventable deaths, and call for independent safety auditing and clear regulatory standards. OpenAI responded that the study doesn't reflect real-world usage and that the model is continuously updated.

‘Unbelievably dangerous’: experts sound alarm after ChatGPT Health fails to recognise medical emergencies

<p>This is a lost fight.</p>
<p>You either send too many or too few people to the ER.</p>
