Experts sound alarm after ChatGPT Health fails to recognise medical emergencies
theguardian.com
February 27, 2026
4 min read
Summary
ChatGPT Health frequently fails to recognize medical emergencies and suicidal ideation, raising concerns among experts about potential harm and death. OpenAI launched the Health feature in January, allowing users to connect medical records and wellness apps for health advice, with over 40 million inquiries reported.
Key Takeaways
- ChatGPT Health fails to recognize medical emergencies, under-triaging over 51% of cases that required immediate hospital care.
- The platform is particularly ineffective at detecting suicidal ideation, raising concerns about its potential to cause harm.
- In a simulated asthma scenario, ChatGPT Health advised waiting for treatment despite identifying early warning signs of respiratory failure.
- The AI was nearly 12 times more likely to downplay symptoms based on user input suggesting they were not serious.
Community Sentiment
NegativePositives
- ChatGPT has provided valuable medical advice for minor issues, demonstrating its potential utility in preliminary health assessments.
- The user's experience highlights that AI can assist in generating hypotheses for medical conditions, which can be beneficial in guiding doctor consultations.
Concerns
- ChatGPT failed to recognize a serious medical issue, leading to a critical situation that required emergency surgery, raising concerns about its reliability in urgent scenarios.
- The study's methodology of testing AI against predetermined outcomes rather than real-world applications limits its relevance, suggesting that AI's performance in actual medical settings remains unverified.
- There are significant reliability issues with AI tools, as they can lead to incorrect information that may result in serious consequences for users.
Related Articles
Marriage over, €100,000 down the drain: the AI users whose lives were wrecked by delusion
Mar 26, 2026
Father claims Google's AI product fuelled son's delusional spiral
Mar 4, 2026
As AI enters the operating room, reports arise of botched surgeries
Feb 9, 2026
Health NZ staff told to stop using ChatGPT to write clinical notes
Mar 25, 2026
Folk are getting dangerously attached to AI that always tells them they're right
Mar 28, 2026
Source
theguardian.com
Published
February 27, 2026
Reading Time
4 minutes
Relevance Score
58/100
Why It Matters
This page is optimized for focused reading: quick context up top, a clean summary block, and a direct path to the original source when you want the full story.