We tend to assume that better access to medical knowledge should make patients feel more certain about their health. But AI may do the opposite.
Medicine contains enormous amounts of uncertainty that patients rarely see. Traditionally, doctors have been the gatekeepers deciding if something is worth worrying about. A headache is probably benign, but there are dangerous causes. An abnormal lab result is probably insignificant, but occasionally it is the first sign of disease. A medication is usually safe, but there are rare complications. Physicians don’t eliminate these possibilities; much of clinical judgment consists of deciding which possibilities are important enough to pursue and which uncertainties are safe to live with. We may consider a dangerous diagnosis, dismiss it after asking a few questions, and move on. And if we disregard it, you may never know about it.
Patients rarely see the dozens of possibilities we considered and decided weren’t worth pursuing. But AI makes those hidden possibilities remarkably easy for anyone to explore. A patient can ask generative AI what else a symptom could represent, what rare diagnoses should be considered, which tests could rule them out, how often those tests miss disease, and what additional testing might provide still more certainty.
That can be enormously valuable. Patients will discover things that clinicians miss, understand their health better, and participate more meaningfully in their care. But there is no natural endpoint to the process. Each answer creates another question, especially when the cost of asking approaches zero.
But MRIs and genetic testing aren’t free. AI makes uncertainty cheap to explore without making it cheap to resolve. Ruling out the next possibility may still require a physician, a specialist, a CT scan, or a biopsy. No amount of tests can reduce the probability to zero, though. In medicine, we often remind ourselves that “low risk is not no risk.” Clinical judgment requires deciding when low risk is low enough.
After automation, the challenge for patients will be learning when to stop exploring unlikely options and listen to their doctors. If they don’t, they’ll walk away with more questions about their health than answers.
