Skip to content
Submit incident
AII-1001-92333
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

Comedian delayed emergency gallbladder surgery after ChatGPT self-diagnosis

Venezuelan comedian George Harris consulted ChatGPT for acute abdominal symptoms while in Madrid, and the reassurance he received delayed his medical evaluation by six days. He eventually required emergency gallbladder surgery in Miami and subsequently warned against relying on AI for medical self-diagnosis of serious symptoms.

26 April 2025 Β· OpenAI Β· Safety and wellbeing
26 April 2025
AII-1001-92333
Quick facts
Category
Safety and wellbeing
Country
Spain
Date
26 April 2025
Drafted by
Raidu Incident Detection

What happened

During a dinner in Madrid, Venezuelan comedian George Harris experienced abdominal bloating and pain and initially considered calling an ambulance, believing he was having a heart attack. Instead, he consulted ChatGPT about his symptoms, which reportedly made him feel better temporarily.

Harris continued his trip to Miami without seeking medical care. Upon arrival, after eating two empanadas, the pain intensified severely. He eventually visited an emergency room where medical tests revealed an inflamed gallbladder requiring immediate surgical removal.

Reported impact

Harris required emergency surgery in Miami to remove his inflamed gallbladder. He subsequently reflected that relying on ChatGPT for self-diagnosis was dangerous and advised others to seek immediate medical attention from healthcare professionals for severe symptoms.

Index analysis

ChatGPT provided reassurance that reduced Harris's concern about his condition, delaying appropriate medical evaluation by approximately six days. The system was used for medical self-diagnosis of acute abdominal symptoms without disclaimers about the need for professional medical evaluation.

Classification

Category
Safety and wellbeing
Industry
Healthcare, drugs and biotechnology
Country
Spain
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Physical injury
Harmed
Consumers
Failure modes
Overreliance, Unsafe recommendation
AI systems
ChatGPT

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident