A patient was troubled for years by unexplained symptoms, undergoing multiple detailed physical examinations and consulting with specialists, but the cause could never be found. Unexpectedly, after submitting his medical reports and detailed symptoms for analysis by ChatGPT, the true cause was identified. Even his attending physician agreed with AI’s analysis, and a targeted treatment brought significant improvement.
An American user shared his experience on Reddit, stating that for over a decade he suffered from numbness in his hands and feet as well as brain fog. During this period, he underwent spinal MRIs, CT scans, blood tests, and other thorough investigations. He even suspected Lyme disease, but all reports came back normal, leaving him ever more confused about his condition.
With no answers in sight, he decided to organize his medical reports, symptom records, and medical history, and hand them over to ChatGPT for analysis. To his surprise, the AI system suggested his symptoms might be related to a genetic mutation—specifically, the A1298C variant in the MTHFR gene.
ChatGPT further pointed out that even if vitamin B12 levels in the blood are normal, metabolic abnormalities in the gene may prevent the body from effectively utilizing these nutrients, leading to neurological discomfort.
The patient then discussed the AI’s analysis with his doctor, who agreed with the AI’s opinion and recommended an MTHFR gene test. The test results confirmed that he indeed carried the mutation.
Based on the information provided by AI and his doctor’s advice, he began supplementing specific nutrients, and the recurring brain fog and limb numbness he had suffered for years significantly improved.
This story sparked widespread attention—even OpenAI’s former president, Greg Brockman, shared the post on X.
He pointed out that AI can integrate MRI, CT, and laboratory data to discover contradictions between symptoms and nutrient absorption and reasonably speculate on possible methylation metabolic disorders. He believes it is time for advanced medical AIs to become a standard “second opinion” in diagnostic processes.
Although such cases of AI-assisted diagnosis are remarkable, the patient specifically cautioned that he never used ChatGPT’s suggestions as his sole basis, but always sought confirmation from professional doctors to avoid misinterpretation from unilateral sources.
Despite generative AI’s promising strengths in data organization and reasoning, the medical community generally believes AI should play a supporting role rather than replace physicians.