Chat-GPT — Told My Patient He Was Dying.

How a mistaken fear led one man to rediscover his family, friendships, and himself.

Photo by Vitaly Gariev on Unsplash

I am a doctor by profession, and every day I meet countless patients.

Some patients come with pain, some with hope, and yes, some simply with the trust that maybe someone will listen to them.

But a few days ago, a patient came to my clinic who taught me more about life than about medicine.

He was a man of about 35 years of age.

The anxiety on his face was clearly visible. As soon as he sat down, he said to me,

“Doctor, I think I am going to die. I think I have some very serious disease.”

I asked him in detail, “First tell me about your problem. What happened?”

He explained that he woke up in the morning, checked his phone, watched some useless videos, had tea, and, as usual, decided that from today he would make his life a little better. But, as happens every time, that “today” was postponed to some other day as well.

Everything was fine until afternoon. Then suddenly there was a slight pain in his chest. Nothing major, just enough to catch his attention. For a few days, he had been feeling tired, and sometimes he would experience mild dizziness.

I tried to ignore it. But then I

opened Chat-GPT and started searching for my symptoms.

Five minutes later, I was convinced that my condition was serious.

At that moment, I felt more fear than pain.

Not the fear that I would die.

Doctor, the fear was that I had not even really started living yet.

I asked him to lie down on the examination table for a check-up, and during the examination, his symptoms did not seem very serious. But the fear on his face was no less than a serious illness.

During our conversation, he had become convinced that he had very little time left.

I prescribed the necessary tests and reassured him that we should wait for the reports first.

But even after he left, his face remained in my mind.

Two days later, I asked him to return with the reports. In the meantime, I gave him some placebo so that he would feel he was at least taking medicine for the illness.

This time, he had the reports in his hand, and there was no sign of the earlier panic on his face.

The reports were almost normal.

There was no serious disease.

Smiling, I said,

“You are absolutely fine.”

Hearing this, relief was clearly visible in his eyes.

But then what he said left me silent for a few moments.

He said,

“Doctor, during these few days when I thought I might be seriously ill, for the first time I truly looked at my life.”

He continued,

“I realised that I had always lived worrying about the future. I became so busy earning money, fulfilling responsibilities, and meeting other people’s expectations that I forgot to live.”

He told me that during those days he had long conversations with his parents.

He called old friends.

For the first time, he thanked his wife without any particular reason.

He spent time with his children.

And he let go of many complaints he had carried in his heart for years.

Then he smiled and said,

“Strangely, I wasn’t actually sick, but during these few days I felt more alive than ever before.”

And laughing, he said,

“Thanks to Chat-GPT, which gave me the wrong answer but taught me the meaning of living.”

After he left, I kept thinking about his words for a long time.

That in the era we live in today, is AI really diagnosing correctly? People are becoming more dependent on machines than on humans.

In a journal published in the British Medical Journal, several major AI chatbots, including ChatGPT, were tested. It was found that more than half (46%) of medical and health-related answers were problematic, incorrect, or incomplete. Chatbots often presented non-scientific claims on the same level as scientific claims, which can mislead patients. — BMJ 2026; 392 doi: https://doi.org/10.1136/bmj.s438 (Published 04 March 2026) Cite this as: BMJ 2026;392:s438

Then there is another one — The Guardian / Mount Sinai. Experts expressed serious concerns about the performance of “ChatGPT Health.” In a test conducted on more than 60 realistic medical scenarios, the AI underestimated (under-triaged) over 50% of emergency cases. For example, even in life-threatening conditions such as diabetes and respiratory failure, the AI advised patients to wait rather than seek medical attention.- https://share.google/kFcJ5KlqNyPC6c7PP

It is true that GPT may be correct in some cases, but when it comes to your health, do not make this mistake.

And there was one more lesson for life as well-

All of us are running in one race or another.

We assume that we have plenty of time.

That is why we delay expressing love.

We delay asking for forgiveness.

We even delay being happy.

But when we feel that time is limited, suddenly every relationship starts to feel important.

The value of every conversation becomes clear.

And every day begins to feel like a gift.

As a doctor, that day I did not just treat a patient.

Rather, a patient reminded me that life’s greatest guarantee is not how much time we have, but how we choose to live the time we do have.

Sometimes it is not illness, but the fear of illness, that teaches us life’s greatest truth.

From a Doctor’s Diary