ChatGPT for Health Advice: Why a Doctor Warns This Could Be Dangerous

In 2024 and 2025, a new pattern became apparent in neurology clinics across India: patients arriving with a diagnosis they had made themselves using ChatGPT — convinced they had a brain tumour, ALS, or MS — based on a list of symptoms they had described to an AI. Some were correct. Most were not. And some had delayed seeking care for a genuinely treatable condition while reassured by an AI that their symptoms were benign.

What ChatGPT and AI Cannot Do

Large language models like ChatGPT generate plausible-sounding medical text. They are trained on medical literature, forums, and patient stories. They are exceptionally good at summarising general medical information. What they cannot do:

• Examine you — feel your abdomen, test your reflexes, assess your gait, look in your eyes
• Review your complete medication list, allergies, and medical history in context
• Order or interpret investigations specific to your case
• Apply clinical judgment that integrates imprecise, incomplete information the way a doctor does
• Take responsibility for a wrong recommendation

Cyberchondria — The New Medical Condition

Cyberchondria is health anxiety amplified by internet and AI symptom searching. The algorithm has a systematic problem: for any symptom combination, the most dangerous possible diagnosis appears in the search results. A headache search returns brain tumour. Chest pain returns heart attack. The result: a cycle of escalating anxiety, more searching, more alarming results, more anxiety — sometimes preventing people from seeing a real doctor because they are too frightened of what they might learn.

When Self-Research Delays Real Care

The most dangerous pattern: a person with stroke symptoms searches "sudden arm weakness" on ChatGPT, receives reassurance that it could be many benign causes, and waits several hours before going to hospital. The 4.5-hour thrombolysis window is missed. This is not hypothetical — it is happening.

The Right Role for AI in Health

AI is genuinely useful for: understanding what a diagnosis means after a doctor has told you, preparing questions to ask your doctor, finding reputable health information. It is not safe as a substitute for diagnosis or for making decisions about whether to seek emergency care.

Sri Anand CNC, Chanda Nagar, Hyderabad. Call +91 90633 66983 — real consultation with real doctors.