A third of American adults now turn to an AI chatbot for at least one health-related task — not as a novelty, but as a real, documented, mainstream behavior. A new Pew Research survey of 3,488 US adults puts a precise number on something that’s been happening quietly for a while: AI chatbots have become real healthcare information infrastructure, without most of the scrutiny that usually accompanies a shift this large.

The real numbers

Pew’s survey found 34% of US adults now use AI chatbots for at least one health-related task. Broken down by what people are actually doing with them: 28% use a chatbot to fetch quick health information, 25% use one to try to understand what’s causing a specific symptom, and 22% use one specifically to access low-cost information they might otherwise have to pay a provider to get. None of these are edge cases — they’re the three most basic reasons anyone interacts with a doctor’s office or a health website, now routed through a chatbot instead, or alongside one.

Adoption isn’t evenly spread — and that unevenness is its own story

The real, notable detail in Pew’s data isn’t just the 34% headline figure — it’s who’s driving it. Adoption skews significantly younger and more diverse than the general population: 56% among Asian American adults and 44% among adults under 30, both well above the 34% overall average. That’s a genuinely different pattern than a lot of technology adoption curves, where usage tends to concentrate among a narrower demographic. Real questions follow directly from that skew: are younger and Asian American adults using chatbots for health information because of real gaps in access to affordable care, language or cultural fit with traditional healthcare settings, or simply higher general comfort with AI tools — and each of those explanations would call for a different policy or product response.

Why this deserves the same scrutiny telehealth got when it scaled

When telehealth platforms and formal health-information websites scaled up over the past two decades, they came with real, substantial debate over accuracy, liability, and equitable access — regulatory review, clinical validation studies, and ongoing scrutiny of whether the information being delivered was actually correct and appropriately cautious. AI chatbots reaching a third of the adult population for health tasks is a comparable scale of adoption, and it’s happening with comparatively little of that same structured scrutiny so far. That’s not necessarily a scandal — general-purpose chatbots have gotten measurably better at hedging medical claims and recommending professional care for serious symptoms — but a third of the population routing real health questions through a system that wasn’t built or regulated as a medical device is worth tracking with the same rigor other health-information channels received.

The genuine upside this data also reveals

It’s worth being fair to the other side of this: 22% of adults using a chatbot specifically to access low-cost health information is a real signal of unmet need in the existing system, not just a novelty behavior. For someone without easy access to a doctor, or facing a cost barrier to a basic question, a chatbot that can explain what a symptom might mean or clarify medical terminology is a genuine improvement over no information at all — the honest framing isn’t “this is dangerous” or “this is great,” it’s that a real behavior shift this large deserves real, ongoing measurement of both the access benefit and the accuracy risk, not a snap judgment in either direction.

The honest takeaway

AI chatbots didn’t replace healthcare — they quietly became a real, load-bearing part of how a third of American adults get basic health information, skewing notably toward younger and Asian American adults. That’s now a real, measurable phenomenon rather than a hypothetical future scenario, and it deserves the same kind of structured, ongoing scrutiny that every other major shift in how people access health information has received.

For the full, regularly updated picture of what else is happening in AI right now, see our daily AI news briefing.

More AI deep dives