AIIMS doctors caution that using ChatGPT for medical advice can be dangerous, after a patient suffered internal bleeding from painkillers suggested by a chatbot.
In a serious warning to the public, Dr. Uma Kumar, Head of Rheumatology at AIIMS New Delhi, cautioned people against using artificial intelligence chatbots for medical self-diagnosis. Speaking to the media after a recent case at the institute, she highlighted the dangers of acting on automated health guidance. ().
The alert followed an incident in which a patient suffered severe internal bleeding after managing back pain based on advice generated by an artificial intelligence chatbot. The individual took non-steroidal anti-inflammatory drugs without consulting a doctor or undergoing basic medical tests.
TOP INSIGHT
Did You Know?
A patient treating back pain with chatbot advice suffered life-threatening internal bleeding after taking painkillers without tests or a doctor’s guidance.
#chatgpt #patientsafety #medindia
When Automated Health Advice Turns Harmful
Doctors at AIIMS revealed that the patient depended on an artificial intelligence tool to address ongoing back pain instead of seeking professional medical care. The chatbot suggested commonly used painkillers, which the patient purchased and consumed independently.
The artificial intelligence system had no knowledge of the patient’s medical history or their vulnerability to stomach and intestinal complications. What seemed like a routine remedy led to a life-threatening episode of internal bleeding.
Why Medical Diagnosis Requires Human Judgment
Physicians note that this reflects a growing trend in which quick online answers are replacing proper medical evaluation, even for medicines that are easily available over the counter.
Dr. Kumar explained that medical diagnosis follows a structured method known as diagnosis by exclusion. Doctors eliminate possible causes through physical examinations, laboratory tests, imaging, and patient history before deciding on treatment.
An artificial intelligence model, by contrast, operates by matching patterns in data. It cannot examine a patient, recognize physical warning signs, or determine whether a symptom signals a deeper condition. In this case, proper investigations would likely have revealed a high risk of bleeding, a step that was completely skipped.
Confident Responses That Can be Incorrect
Medical professionals are increasingly worried about what are often referred to as artificial intelligence hallucinations, where chatbots deliver information confidently despite gaps or errors.
Although platforms such as ChatGPT include disclaimers, their tone can sound authoritative, especially to someone experiencing pain. Advising the use of non-steroidal anti-inflammatory drugs is not unusual in general practice, but for this patient, it proved dangerous.
Without a doctor to assess contraindications or hidden conditions, even a common recommendation can result in serious harm.
Calls for Public Awareness and Oversight
The episode has reignited debate over how artificial intelligence platforms should respond to health-related questions. Doctors at AIIMS are urging people to treat online tools as sources of general information, not as personal treatment guides.
Experts agree that artificial intelligence can support healthcare in limited areas such as administrative work or research assistance, but it should never replace professional diagnosis or supervision.
There are also demands for stronger public awareness and clearer regulation to prevent similar incidents. Doctors continue to emphasize that medical judgment, grounded in examination and evidence, cannot be replaced by algorithms.
References:
- ‘Bleeding after self-diagnosis’: AIIMS doctor flags risks of using ChatGPT for health- (https://www.hindustantimes.com/india-news/bleeding-after-self-diagnosis-aiims-doctor-flags-risks-of-using-chatgpt-for-health-101768556393452.html )
- AIIMS Doctor Issues Warning After Patient Follows ChatGPT Advice, Suffers Internal Bleeding – ( https://www.healthandme.com/health-news/aiims-doctor-issues-warning-after-patient-follows-chatgpt-advice-suffers-internal-bleeding-article-153479571 )
Source-Medindia