Pew 2026: Half of Americans Now Use AI Chatbots

Pew Research Center’s “Americans and AI 2026” study, released June 17, 2026, marks a quiet tipping point: 49% of US adults now use AI chatbots, up from 33% in 2024 and 23% in 2023. Half the country has adopted a technology that barely existed three years ago — and ChatGPT alone now reaches 44% of adults.

What makes the report consequential for marketers is not the adoption number on its own. It is the second number underneath it: 60% of US adults now read AI-generated summaries in their search results. Discovery is fragmenting across three surfaces at once — traditional search, AI summaries, and direct chatbot queries — while trust in those surfaces lags well behind usage.

This guide works through the data that matters, separates the primary Pew findings from the secondary coverage, and translates the numbers into concrete implications for how brands earn visibility and credibility in an AI-mediated discovery landscape. Every figure is attributed; where a statistic comes only from secondary reporting, we say so.

Pew’s headline finding is a clean trend line. As of early 2026, 49% of US adults report using AI chatbots — up from 33% in summer 2024 and 23% in 2023, the first year Pew tracked the metric. That is a 16-point jump in two years, the kind of curve that takes a technology from “early adopters” to “most of your audience.”

Intensity is climbing alongside reach. A full 24% of US adults now use AI chatbots daily — 12% several times a day and 4% describe it as “almost constantly.” Another 25% use them several times a week or less. For a marketer, the practical reading is simple: chatbot use is no longer an occasional novelty for a sliver of your market. For a quarter of US adults it is a daily habit.

Independent surveys point the same direction. A separate SSRS panel fielded in February 2026 found 52% of Americans using AI platforms on a weekly basis, and a 2025 Brookings analysis found 57% had used generative AI for personal purposes — with a steep education gradient (67% among adults with a bachelor’s degree or higher). Different instruments, different definitions, same conclusion: adoption has crossed into the mainstream.

ChatGPT remains the default. Pew reports 44% of US adults have used it — more than doubling from 18% in 2023 (a 2.4x increase, which is why “more than doubled” is the precise phrasing). It sits well ahead of the field: Gemini at 24%, Copilot at 17%, Meta AI at 14%, Grok at 8%, Claude at 6%, and Character.ai at 3%.

For marketers, the long tail matters as much as the leader. Brand presence inside AI answers is no longer a single-platform problem — it is a portfolio. The cohort reaching for Gemini is largely inside Google’s ecosystem; the Copilot cohort skews toward Microsoft and workplace tools; Meta AI surfaces inside social apps people already open daily. Optimizing for one assistant no longer covers the audience.

One caveat worth keeping straight: Pew measures adult adoption of each platform, not the share of total AI queries each handles. A platform can reach a wide slice of adults while handling a smaller slice of queries, or vice versa. For the query-share angle — and the running debate about whether any single assistant still commands a majority of AI search — see our breakdown of ChatGPT’s platform market share below 50%, which uses query and traffic data rather than self-reported adoption.

The single most marketing-relevant number in the study: 60% of US adults report reading AI-generated summaries in their search results, such as Google’s AI Overviews. Thirty percent have not, and 10% are unsure. Men read them slightly more than women (63% versus 57%).

The careful reading matters here. Reading an AI summary is not the same as skipping the organic results beneath it — a person can read the summary and still click through. What the figure does establish is that for a majority of adults, an AI-generated layer now sits above the traditional blue links, framing the answer before they ever reach a website. Your content is increasingly being read in paraphrase before it is read directly.

That has a direct consequence for how content earns visibility. The page that gets summarized accurately — and attributed — wins mind-share even when the click never happens. The page that gets summarized inaccurately, or omitted, loses ground it used to hold through ranking alone. This is the practical core of generative engine optimization, and it is why our work on GEO content strategy for agencies now sits alongside traditional SEO rather than after it.

The traffic consequence of that summary layer is already showing up in referral data. For how AI summaries and chatbots are translating into measurable referral visits — and how that splits across Gemini, ChatGPT, and the rest — see our AI referral traffic breakdown by platform, which pairs naturally with Pew’s consumer-attitude data.

Here is the paradox that defines the report. Half of Americans use AI chatbots, but only 29% of those users say they have “a lot” or “some” trust in the information chatbots provide — meaning roughly seven in ten chatbot users hold little or no trust in what they read. People are using a tool they do not believe. (Do not invert this: it is the usage rate that is high, not the trust rate.)

Secondary coverage of the report attributes that distrust to three main concerns among the skeptical: the spread of misinformation, a lack of transparency about how answers are generated, and the potential for bias or harmful content. We have seen those specific percentage breakdowns cited as roughly three-quarters and two-thirds, but they appear in secondary reporting rather than a directly quotable Pew table — so we frame them qualitatively rather than printing precise figures we cannot anchor to the primary source.

The interesting move for marketers is to stop treating trust as one number. Pew reports a single 29% figure, but trust clearly depends on what the tool is being used for. Below is our synthesis: Pew’s use-case shares mapped against the trust risk each carries and the content response it implies. The trust and risk columns are our editorial reading of Pew’s overall distrust signal applied per use-case — not Pew-reported trust scores.

The insight jumps out once the columns sit side by side: medical advice is used by one in five chatbot users yet sits in the highest-risk trust band. For any brand in health, finance, or law, that combination is the warning. Audiences are bringing consequential questions to tools they distrust — which makes explicit human-verification messaging and visible sourcing a competitive advantage, not a compliance chore.

The mood underneath the adoption curve is wary. 63% of Americans say AI is advancing too quickly; only 19% say the pace is about right and 2% say too slowly. Looking 20 years out, just 16% predict AI will have a net-positive impact on society, while 40% predict net-negative and 31% expect it to be roughly balanced.

Concern is specific, not vague. 71% of Americans believe AI will make personal information less secure. And confidence in oversight is thin on both sides of the market: 67% have little or no confidence in elected officials to regulate AI effectively (up from 62% in 2024), and 59% distrust US companies to deploy AI responsibly — unchanged from 2024. People are skeptical of both the regulators and the builders.

For brands, that environment rewards restraint over hype. The audience is not waiting to be told AI is exciting; a clear majority already thinks it is moving too fast. Copy that leads with transparency, data handling, and human oversight speaks to where people actually are. Copy that leads with magic and inevitability argues against the prevailing mood.

Conventional marketing wisdom says younger audiences are the AI enthusiasts — the ones who respond to AI-as-exciting framing. Pew inverts it. Yes, adoption follows the expected age gradient: 66% of 18–29 year-olds use chatbots, falling to 61% for 30–49, 42% for 50–64, and 23% for those 65 and older. The young are the heavy users.

But they are also the pessimists. 18–29 year-olds are more likely than their elders to predict a negative societal impact (48% versus 37% for those 50 and older) and a negative personal impact (37% versus 28%). They use AI more and believe in it less. That is the single most counter-intuitive finding in the report for anyone building a marketing message.

The implication is direct: do not assume Gen Z and younger Millennials are AI boosters waiting for an enthusiastic pitch. They are sophisticated, high-frequency users who are also clear-eyed about the downsides. With this audience, transparency and control messaging — what data you take, what the AI can and cannot do, how a human stays in the loop — is more likely to land than novelty or awe.

Stack the two headline numbers and a structural shift becomes visible. If 60% of US adults read AI summaries and 49% actively use chatbots, then every piece of branded content now faces at least three distinct discovery surfaces: the traditional search results page, the AI summary layered above it, and the direct chatbot query that never touches a results page at all.

Most marketing organizations still treat these as one funnel — the SEO funnel — with AI bolted on. The data argues for treating them as three surfaces with different optimization logic, different trust profiles, and different measurement. A page can rank well, get summarized poorly, and be omitted from chatbot answers entirely. All three need to be working for a brand to hold its visibility.

For the quantitative companion to this picture — search-market and AI-engine share figures that sit alongside Pew’s consumer-attitude data — see our running AI search market share statistics. Pew tells you how people feel and what they use; the market-share data tells you where the queries are actually going.

The Pew data does not call for a new channel; it calls for a posture. Adoption is mainstream, trust is low, and discovery is fragmented. The brands that benefit will be the ones that treat AI-mediated discovery as a credibility problem first and a visibility problem second.

Operationally, this is where measurement has to expand. Tracking only organic rank misses two of the three surfaces. The brands getting ahead are instrumenting how often they appear in AI summaries, which chatbots cite them, and whether the paraphrase is accurate — then feeding that back into content structure. That cross-surface discipline is the heart of our agentic SEO and GEO service, and it sits alongside the broader AI transformation work that helps teams adopt these tools without losing the trust their audience is already short on.

The forward view: as AI summaries become the default reading layer and chatbot use deepens from weekly to daily, the gap between brands that are legible to machines and brands that are not will widen faster than the old SEO gap ever did. The advantage is available now, while most competitors are still treating AI discovery as an experiment rather than the surface where half their audience already lives.