From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health is a Pew Research Center report published on 25 August 2026 by Emma Kikuchi, Giancarlo Pasquini and Eileen Yam. It draws on a survey of 3,488 U.S. adults conducted 22–28 June 2026 through the Center's American Trends Panel, and reports both self-reported use of AI chatbots for health purposes and public views on chatbot use for social and mental health support (Source: pewresearch.org).
The report's two chapters produce findings that run in different directions: users of chatbots for health rate the information they receive highly, while the general public is more likely to say chatbot use for loneliness, depression or stress does more harm than good.
Use for health information
Pew defines "chatbot health users" as adults who ever use AI chatbots for at least one of eight health and medical reasons. That group is 34% of U.S. adults. Nearly all cite multiple reasons.
| Reason for use | % of U.S. adults |
|---|---|
| To get health information quickly | 28 |
| To figure out what's causing symptoms | 25 |
| To get health information at little to no cost | 22 |
| To get health information about treatments | 22 |
| To learn more about a diagnosis from a doctor | 22 |
| To help understand lab results | 20 |
| To get health information they are uncomfortable talking about | 18 |
| To help decide whether to go to a doctor | 15 |
| Any of the eight reasons | 34 |
The report groups the motivations as convenience (28% for speed), diagnostic (25% for symptom causes), comprehension of information obtained elsewhere (20–22% for lab results, treatments and a doctor's diagnosis), and cost (22%).
Demographic patterns
Chatbot health use tracks overall chatbot use. It is highest among Asian adults at 56% (an estimate Pew notes is representative of English speakers only), against 39% of Hispanic adults, 35% of Black adults and 29% of White adults. It falls with age — 44% of adults 18–29, 40% of those 30–49, 32% of those 50–64 and 17% of those 65 and older — and rises with education (44% of four-year college graduates against 23% of those with a high school education or less) and income (48% upper-income against 28% lower-income).
The sharpest split is by disposition toward AI generally. Among adults who say increased AI use makes them feel more excited, 60% are chatbot health users; among those who say it makes them more concerned, 21% are.
Perceived helpfulness and comfort with data
Among chatbot health users, 47% rate the health information they get as extremely or very helpful, 48% as somewhat helpful, and 5% as not too or not at all helpful.
Comfort with sharing personal health information is more evenly divided: 29% of chatbot health users say they would be extremely or very comfortable, 42% somewhat comfortable, and 26% not too or not at all comfortable. The report places this alongside scrutiny of platforms for storing and selling health data or using it to train large language models, and the concurrent release of healthcare chatbot tools by technology companies.
Views on social and mental health use
Asked whether chatbots do more to help or hurt people who use them for support, U.S. adults lean negative on all three items surveyed, with large uncertain shares.
| Use case | Hurt | Help | Neither | Not sure |
|---|---|---|---|---|
| Loneliness | 39 | 19 | 14 | 27 |
| Depression | 36 | 17 | 14 | 33 |
| Stress | 29 | 22 | 18 | 31 |
Roughly three-in-ten say they are not sure on each item, and a further 14–18% say chatbots neither help nor hurt.
The age pattern runs opposite to the usage pattern. Adults under 30 use chatbots for emotional support more than older adults do, and are also the most negative about the practice: 49% say chatbots do more to hurt people who use them for loneliness and 41% say the same for depression, against 23% saying they help on each. Uncertainty rises steeply with age — among adults 65 and older, 40% are not sure on loneliness, 46% on depression and 44% on stress. Stress is the one item where adults under 50 are about as likely to say chatbots help as hurt.
Asian adults are the only racial or ethnic group that skews positive on any item, saying chatbots do more to help than hurt on loneliness (37% against 25%) and stress (34% against 21%), and splitting on depression (30% against 26%). White adults are the most negative on loneliness (43% hurt against 16% help) and depression (39% against 14%). Pew reports modest or no differences by gender, education or income on these items.
Provenance
The survey was fielded 22–28 June 2026 on the American Trends Panel with 3,488 respondents. The report, questionnaire and topline are published as PDFs alongside the web version. The report builds on the Center's earlier work on Americans' health information sources and on its June 2026 report Americans and AI 2026, which supplies the underlying chatbot-use base rates. Chapters 3 (acknowledgments) and 4 (methodology) were not captured in the vault's copy; weighting and margin-of-error detail is published separately at pewresearch.org.
Relationships
- supports: Healthcare — AI Deployment — consumer-side adoption figures for AI in health.
- supports: AI Mental Health and Psychological Harm — public opinion on chatbot use for loneliness, depression and stress.
- related: Pew Research Center, AI Companions, Public Opinion on AI, AI and Privacy