Industry report · October 2026
AI assistants and consumer health shopping in the US, October 2026
Who in the US asks AI assistants about health, what they ask, what happens to the click, and which brands the answers name. Every figure carries its source, its date and what it measures.
Published October 18, 2026, sources re-read the same day.
Summary
About half of US adults use an AI chatbot (51%, up from 49% in February), and about a third use one for at least one of eight health reasons (34%).
Source: Pew Research Center, survey of 3,488 US adults, June 22-28, 2026 (Wave 195 topline; health report).
Regular use for health has nearly doubled in two years: 29% of adults use AI tools or chatbots for health information at least monthly, against 17% in June 2024.
Source: KFF, survey of 2,480 US adults, May 7-31, 2026.
ChatGPT leads and Gemini is second in both national surveys: 44% and 24% of adults ever use them, and 31% and 21% used them in the past week.
Source: Pew Research Center, survey of 5,119 US adults, February 17-23, 2026. Epoch AI / Ipsos, survey of 2,021 US adults, March 3-6, 2026.
The biggest single health intent is plain explanation, such as how a medication works: 40.8% of Copilot health conversations in a global study.
Source: Nature Health (Costa-Gomes et al.), 617,827 Copilot health conversations, about 22% from the US, January 2026. Global figure, not US only. Licensed CC BY 4.0.
Google users clicked a search result on 8% of visits to pages with an AI summary, against 15% on pages without one.
Source: Pew Research Center, browsing data from 900 US adults, March 2025.
Among US shoppers, 45% use AI chatbots or assistants for at least some of their beauty and personal care purchases.
Source: Criteo, survey of 6,307 shoppers in six countries, US figures, June 2026. Criteo sells retail advertising.
In our own counts, how concentrated the answers are depends on the category: every one of 120 pain relief answers named a brand, while 26 of 60 sleep supplement answers stopped at an ingredient.
Source: Supracited census of Claude with web search, 240 answers across three categories, collected October 17, 2026. One assistant, wide ranges; see below.
Findings 1 and 2 answer different questions (Pew asks whether, KFF how often), so they sit side by side in words rather than on one chart.
Every source we read, the figures we kept off the charts, and our category counts in detail.
ChatGPT leads, Gemini is second, and the gap depends on the ruler
Pew asked whether people ever use each chatbot; Ipsos asked about the past week, in early March. Both put the assistants in the same order.
Share of US adults who use each AI assistant
- Ever use (Pew, February 2026)
- Used in the past week (Ipsos, March 2026)
| Item | Ever use, Pew, February 2026 | Past week, Ipsos, March 2026 |
|---|---|---|
| ChatGPT | 44% | 31% |
| Gemini | 24% | 21% |
| Copilot | 17% | 11% |
| Meta AI | 14% | 8% |
| Grok | 8% | 5% |
| Claude | 6% | 3% |
| Perplexity | not asked | 2% |
Source: Pew Research Center, survey of 5,119 US adults, February 17-23, 2026. Measured: say they ever use each AI chatbot.
Source: Epoch AI / Ipsos, survey of 2,021 US adults 18 and over, Ipsos KnowledgePanel, March 3-6, 2026. Measured: say they used each AI service in the past week. Drawn by Supracited from the published numbers. Two surveys, two time windows: compare the order, not the gap between bars.
Part of what surveys count as Gemini is Google Search: of adults who used Gemini in the past week, 47% counted an AI summary at the top of Google Search results, more than the 39% who opened the Gemini app or site.
Source: Epoch AI / Ipsos, March 3-6, 2026, the same survey as the chart. Measured: how past-week Gemini users used it, several answers allowed; base 264 past-week Gemini users, unweighted.
Commercial panels that count desktop visitors (Comscore), app users (Apptopia) or a share whose method is not documented (StatCounter) also put ChatGPT first. Its lead changes with what each one counts, so their numbers stay off this chart.
Younger adults ask most, and many ask for someone else
Use falls with age, though even among adults 65 and over about one in six has used a chatbot for a health reason.
Share who use AI chatbots for health, by age
- All US adults
- By age
| Item | Share |
|---|---|
| All US adults | 34% |
| Ages 18-29 | 44% |
| Ages 30-49 | 40% |
| Ages 50-64 | 32% |
| Ages 65 and over | 17% |
Source: Pew Research Center, survey of 3,488 US adults, June 22-28, 2026. Measured: ever use AI chatbots for at least one of eight health reasons Pew asked about. Drawn by Supracited from the published numbers.
1 in 7
Symptom and condition conversations about someone else
A child, an aging parent or a partner: 14.5% of symptom conversations and 14.9% of condition conversations.
Source: Nature Health (Costa-Gomes et al.), 617,827 health conversations with Copilot from consumer accounts, worldwide, about 22% from the US, January 2026. Global figure, not US only. Measured: share of Copilot health conversations in each intent; share of symptom and condition conversations about someone else. Licensed CC BY 4.0.
For a brand, the person asking is not always the person who will take the product.
Symptoms and treatments lead, and plain explanation is the biggest intent
Asked about eight health reasons, US adults put speed first, then symptoms, cost, treatments and diagnoses.
Why US adults use AI chatbots for health
| Item | Pew wording | Share |
|---|---|---|
| Get health information quickly | To get health information quickly | 28% |
| Figure out what is causing symptoms | To figure out what's causing symptoms | 25% |
| Information at little or no cost | To get health information at little to no cost | 22% |
| Information about treatments | To get health information about treatments | 22% |
| Learn more about a diagnosis | To learn more about a diagnosis from a doctor | 22% |
| Understand lab results | To help understand lab results | 20% |
| Topics hard to raise in person | To get health information you are uncomfortable talking about | 18% |
| Decide whether to see a doctor | To help decide whether to go to a doctor | 15% |
Source: Pew Research Center, survey of 3,488 US adults, June 22-28, 2026. Measured: say they ever use AI chatbots for each health reason. Drawn by Supracited; reason labels shortened, Pew’s wording is in the data table.
Conversation logs point the same way: in the global Copilot study cited in the summary, the largest share of health conversations asks for an explanation, such as how a medication works, not for a product. Most health questions are not shopping questions, which is worth remembering when reading any count of brand mentions, ours included.
AI summaries take clicks, and shoppers use AI to choose
Pew’s browsing panel shows the click moving into the answer.
What Google users did on a search page
- Page with an AI summary
- Page without one
| Item | Page with an AI summary | Page without one |
|---|---|---|
| Clicked a search result | 8% | 15% |
| Ended the browsing session | 26% | 16% |
| Clicked a link in the summary | 1% | not applicable |
Source: Pew Research Center, browsing data from 900 US adults in Ipsos KnowledgePanel Digital, March 2025. Measured: share of Google search visits on which the user clicked a traditional result, clicked a link inside the AI summary, or ended the browsing session, by whether the page showed an AI summary. Drawn by Supracited from the published numbers.
In personal care the answer already shapes the purchase: among US shoppers who use AI, 59% say products surfaced in AI answers would be highly or moderately influential.
Source: Criteo, US shoppers in a six-country shopper survey (N=6,307 across the US, UK, France, Germany, Japan and South Korea; US subsample size not published), June 2026. Measured: use AI chatbots or assistants for all, most or some of their beauty and personal care purchases; among US shoppers who use AI, say products surfaced in AI answers would be highly or moderately influential. Criteo sells retail advertising.
For OTC medicines and OTC hearing aids we found no published survey of AI shopping.
What our own measurement adds
No source we read counts which brands the answers name for OTC medicines, supplements or hearing aids with a disclosed method. Our census does. So far its health categories rest on one assistant, Claude with web search, so the ranges are wide. Read them as the shape of each category, not as settled shares.
The most-named brand in each category
- Share of answers that named it
- 95% range
| Item | Brand | Answers that named it | Share | 95% range |
|---|---|---|---|---|
| Pain relief and headache | Tylenol | 98 of 120 | 82% | 48 to 100% |
| OTC hearing aids | Jabra Enhance | 45 of 60 | 75% | 42 to 100% |
| Sleep supplements | Thorne | 18 of 60 | 30% | 3 to 62% |
Source: Supracited census of Claude with web search, collected October 17, 2026. Six buyer scenarios per category, asked 20 times each for pain relief and headache; 10 times each for OTC hearing aids and sleep supplements. Measured: share of answers that named the brand. Ranges are 95% cluster-bootstrap ranges, which allow for the same question being asked again.
Hearing aid answers spread across several brands: Eargo came up in 38 of 60 and Lexie in 35, behind Jabra Enhance. Concentration is a property of the category, not a law of AI answers.
Brand names here are measurement results, not recommendations. Full tables: pain relief, OTC hearing aids, sleep supplements. How we count.
Method and sources
Every outside figure on this page was re-read at its live source on October 18, 2026, and figures we could not re-read that way are left out. We chart a figure only where the source’s terms let a commercial site redraw it (Pew Research Center, Epoch AI / Ipsos and the open-access Nature Health article); KFF and Criteo appear in prose. Each chart is our own drawing of published numbers, and none of the sources endorses Supracited or bears responsibility for our reading of their data.
Every figure, with its full citationlicence, how it is used and when it was re-read
Company statements that disclose no sample are left out too. Survey figures describe US adults unless marked global. Our own counts follow the rules on the method page.
-
Chatbot use, US adultsShown as a figure
Pew Research Center, American Trends Panel Wave 195 topline, survey of 3,488 U.S. adults conducted June 22-28, 2026, published August 25, 2026. https://www.pewresearch.org/wp-content/uploads/sites/20/2026/08/PS_2026.8.25_AI-chatbots-health_topline.pdf (accessed October 18, 2026)
-
Ever use of each AI chatbotCharted
Pew Research Center, June 2026, "Americans and AI 2026: Chatbots, Smart Devices and Views on Impact." Survey of 5,119 U.S. adults conducted Feb. 17-23, 2026. https://www.pewresearch.org/chart/americans-report-using-chatgpt-far-more-than-other-chatbots/ (accessed October 18, 2026)
-
AI assistants used in the past weekCharted
Epoch AI / Ipsos National AI Usage Survey, conducted by Ipsos using KnowledgePanel, March 3-6, 2026, 2,021 U.S. adults ages 18+, margin of error plus or minus 2.2 percentage points. https://www.ipsos.com/sites/default/files/ct/news/documents/2026-06/Topline_Epoch%20AI%20-%20Ipsos%20National%20AI%20Usage%20Survey.pdf (accessed October 18, 2026)
-
Chatbot health users, by ageCharted
Pew Research Center, August 2026, "From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health." Survey of 3,488 U.S. adults conducted June 22-28, 2026. https://www.pewresearch.org/chart/asian-americans-and-younger-adults-are-among-the-most-likely-to-be-chatbot-health-users/ (accessed October 18, 2026)
-
Monthly use of AI for health informationCited in prose
KFF Tracking Poll on Health Information and Trust: Use of Social Media and AI For Health Information and Advice, (KFF, June 17, 2026) https://www.kff.org/public-opinion/kff-tracking-poll-on-health-information-and-trust-use-of-social-media-and-ai-for-health-information-and-advice/ (accessed October 18, 2026)
-
Health reasons for using chatbotsCharted
Pew Research Center, August 2026, "From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health." Survey of 3,488 U.S. adults conducted June 22-28, 2026. https://www.pewresearch.org/chart/reasons-for-using-chatbots-for-health-span-convenience-cost-and-medical-advice/ (accessed October 18, 2026)
-
Health intents in Copilot conversationsShown as a figure
Costa-Gomes B, Tolmachev P, Taysom E, et al. "Public use of a generalist LLM chatbot for health queries." Nature Health, published April 16, 2026. Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). https://www.nature.com/articles/s44360-026-00117-x (accessed October 18, 2026)
-
Clicks on Google pages with AI summariesCharted
Pew Research Center, July 2025, "Do people click on links in Google AI summaries?" Browsing data from 900 U.S. adults in Ipsos KnowledgePanel Digital, March 2025. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/ (accessed October 18, 2026)
-
AI use in beauty and personal care shoppingCited in prose
Criteo, "2026 Global Health & Beauty Pulse," Criteo Shopper Survey, June 2026, N=6,307 shoppers in the US, UK, France, Germany, Japan and South Korea. https://www.criteo.com/wp-content/uploads/2026/06/2026-Global-Health-Beauty-Pulse.pdf (accessed October 18, 2026)
The full report goes further
It adds every source we read, the panel and company figures we kept off the charts, the licence terms behind each one, and our category counts in detail.