About one in three U.S. adults have turned to artificial-intelligence chatbots for health information in the past year — roughly the same share who use social media for that purpose — according to a poll published Tuesday by the health-research organization KFF. The same survey found that adults who use AI for health information most frequently were more likely than non-users to believe debunked claims about vaccines.

Among adults who use AI for health information at least weekly, 35% said it was “probably” or “definitely true” that the measles, mumps and rubella (MMR) vaccine causes autism — a claim scientific evidence does not support — compared with 20% of those who never use AI for health advice. KFF said the link held even after accounting for age, race and ethnicity, education and partisanship.

The KFF Tracking Poll on Health Information and Trust surveyed a nationally representative sample of 2,480 U.S. adults from May 7 to 31, 2026, online and by telephone.

Correlation, not proof

The findings describe an association, not a cause. The poll does not show that chatbots instil false beliefs, and it does not capture what answers respondents actually received from AI tools. People already inclined to distrust mainstream medical advice may also be more likely to seek out alternatives, including AI.

A trust problem, not just a tech problem

The poll also points to broad public uncertainty about AI as a health source. KFF found only about a third of adults are confident the health information they encounter on chatbots is accurate, and a majority say they are not confident they can tell what is true from what is false in AI responses — even as adults under 30 use the tools more than any other age group.

The results land amid declining trust in health institutions. KFF has reported that public confidence in the U.S. Centers for Disease Control and Prevention is around its lowest since the COVID-19 pandemic, and that fewer than half of adults are confident in federal agencies’ vaccine recommendations, in the context of recent changes to childhood-vaccination guidance.

What's next

The data sharpens a question facing regulators and AI developers: how to make fast-growing health chatbots reliable for a public that increasingly uses them but struggles to judge their accuracy. With AI now rivaling social media as a health-information channel, the stakes of getting medical answers right are likely to keep rising.