As more people use artificial intelligence to ask about their health, a new survey raises a pointed question about what they are taking away from it.
What the poll found
The KFF Tracking Poll on Health Information and Trust, a survey of about 2,480 U.S. adults conducted in May, found an association between regular use of AI for health information and belief in vaccine falsehoods. Among adults who said they turn to AI for health advice at least weekly, 35 percent said it was probably or definitely true that the MMR vaccine causes autism — a claim that has been thoroughly discredited. Among those who never use AI for health information, 20 percent said the same. Similar gaps appeared on other false statements the poll tested. The link held even after researchers accounted for differences in age, race, education and political affiliation, the Guardian reported.
A link, not a cause
KFF was careful to stress what the finding does — and does not — show. It is a correlation, not proof of cause and effect. The poll cannot say whether using AI chatbots leads people to adopt false beliefs, or whether people already distrustful of mainstream health sources are simply more likely to rely on chatbots in the first place. Both could be true at once, and the survey is a snapshot of one moment, not the final word.
Why there is concern
The worry is grounded in known weaknesses of the technology. AI chatbots can produce confident, authoritative-sounding answers that are wrong — so-called "hallucinations" — sometimes inventing studies or blending facts with fabrication. A patient-safety organization named the misuse of AI chatbots a leading health-technology hazard for 2026, and independent analyses have found that a substantial share of chatbots' answers to medical questions contain problems. Because a chatbot's reply can feel personalized and definitive, users may accept it without the skepticism they might apply to a stranger's post online.
What is being done
Major AI companies have added disclaimers to their health-related features urging users not to treat chatbots as a substitute for professional care, and say they are working to curb confident misinformation. Public-health bodies, including the World Health Organization and the U.S. Centers for Disease Control and Prevention, advise people to check health information against authoritative sources and to consult their own doctors. General-purpose chatbots, regulators note, are not approved medical devices.
What the science says
On the substance, the scientific consensus is not in doubt. Health authorities including the WHO and the CDC hold that approved vaccines are rigorously tested before authorization and monitored for safety afterward, and that they are safe and effective at preventing serious disease. The specific claim that the MMR vaccine causes autism traces back to a fraudulent, long-retracted study and has been contradicted by large bodies of subsequent research.
With roughly a third of Americans now reporting that they have used AI for health information, the poll adds to a growing unease about a tool that can be genuinely useful but also, at times, quietly wrong.
