Medical professionals warn of incorrect ChatGPT diagnoses circulating online
Spanish fact-checkers and Portuguese medical professionals warned that ChatGPT and similar AI chatbots produce incorrect medical diagnoses, after claims circulated online suggesting the tools could detect diseases like cancer. Medical experts stated the systems lack clinical reasoning, comprehensive patient assessment, and access to medical history needed for accurate diagnosis, and can generate plausible-sounding but incorrect recommendations that pose risks to patient health.
What happened
The Spanish fact-checking platform Maldita alerted that ChatGPT and similar AI chatbots can produce incorrect medical diagnoses and do not replace doctors, after claims circulated on social media and websites that the tools successfully diagnosed conditions including skin cancer. Claims such as 'I asked ChatGPT to read my hand and it detected skin cancer' and 'ChatGPT diagnosed a woman with cancer a year before doctors' appeared online, with some testimonials asserting the AI tool is capable of providing medical diagnoses.
According to Maldita and medical professionals quoted by the platform, AI chatbot results are based on probabilities, training data may be unreliable, and the tools cannot perform the comprehensive analysis that a doctor conducts. A doctor cited by Maldita stated that a chatbot cannot perform necessary medical analysis or access sufficient information to make a diagnosis, and does not account for other factors such as comorbidities or patient context.
Reported impact
Medical professionals expressed concern that generative AI systems without quality control or scientific validation can generate clinically incorrect, biased, or decontextualized content, amplifying medical misinformation. The Portuguese Medical Association noted that unreliable AI recommendations appear credible but can be incorrect, and such misinformation compromises health literacy and undermines trust in medical expertise, potentially inducing risky patient behavior.
Index analysis
AI chatbots lack the ability to perform comprehensive patient assessment, do not have access to individual medical history or comorbidity information, and produce outputs based on probabilistic pattern matching rather than clinical reasoning. The tools may generate plausible-sounding but factually incorrect medical recommendations, particularly when training data is unreliable or incomplete.
Governance controls
The Portuguese Medical Association's president called for urgent creation of a national AI agency for healthcare with autonomy, multidisciplinary expertise, and technical competence to monitor and regulate AI use in clinical settings. This proposed entity would function as a permanent oversight structure to protect fundamental medical values including patient centrality, respect for the medical act, doctor-patient trust, and humanized healthcare, and to ensure doctors validate clinical information and medical authority is not diluted by opaque automated systems.
Classification
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.