Summary: Pew Research Center’s March 12, 2026 roundup of five years of surveys found 50% of adults in the United States more concerned than excited about AI, alongside rising teen chatbot use and growing workplace exposure. The time depth makes the trust deficit a durable design constraint, while the results also split by age, topic, and political affiliation rather than describing one market.
The five-year roundup of American views on artificial intelligence arrived from Pew Research Center on 12 March 2026 — researchers Michelle Faverio and Emma Kikuchi condensed its tracking surveys since 2021 into 13 findings in an update to the same series from November 2023. The number that anchors the whole package: 50% of U.S. adults say the expanding role of AI in daily life leaves them more concerned than excited, while only 10% feel more excited than concerned and 38% feel both at once.
What gives the roundup its weight is time depth. A monthly poll reads a mood; five straight years of data read a direction, and the direction is unmistakable — familiarity and usage keep climbing, concern climbs right beside them, and excitement never closes the gap. For any team planning to blanket its product with AI, this set of numbers is the most direct market thermometer on offer.
Five Years, 13 More Points of Concern
The question debuted in Pew’s surveys in 2021, when 37% of U.S. adults were more concerned than excited; the June 2025 wave put the figure at 50%, and American worry also tops most of the other 24 countries Pew has polled. Expectations for society run just as cold: close to half of adults think AI will dull creative thinking and make meaningful relationships harder to build. The scarce good news sits in medicine — 44% expect a positive impact on medical care against 19% who expect harm — while education (24% positive) and jobs (23% positive) trail well behind, with another 16% to 30% unsure depending on the area. Familiarity, for its part, has leapt: 47% of adults have heard “a lot” about AI, 21 points above the 26% of 2022 — and none of that extra knowledge has made anyone calmer.
Teens and Younger Kids Went In Early
In the fall 2025 survey, 64% of U.S. teens ages 13 to 17 had used an AI chatbot. Over half reach for chatbots when searching for information or getting help with schoolwork, and about one in ten hand all or most of their schoolwork over entirely. By purpose: summarising text or video 42%, creating or editing images and video 38%, casual conversation 16%, emotional support or advice 12%. Chatbot cheating is part of the picture — roughly six in ten teens say it happens at least somewhat often at their school, and about a third rate it extremely or very often. The entry point keeps getting younger: parents report 15% chatbot use among 11- to 12-year-olds, voice-assistant use of 53% in that same age band and 12% even below age two, plus smartwatch adoption of 21% among parents of 11- to 12-year-olds and 11% among all parents of children 12 or under.
On the Job: Slow, Steady Adoption
As of September 2025, 21% of U.S. workers said AI covers at least some of their work, up from 16% in 2024; conversely, 65% use it little or not at all. Everyday contact is thickening too: 31% of adults now interact with AI several times a day, against 22% in February 2024. The age divide jumps out — 38% of employed adults under 30 use ChatGPT at work, against 30% for ages 30 to 49 and 18% for those 50 and older — while roughly half of adults under 50 interact with AI daily or more. Simply put, AI already runs as background infrastructure for half of Americans under 50, yet formal adoption inside job workflows still lags well behind everyday personal use.
Where Experts and the Public Diverge
Pew’s 2024 expert comparison supplies the sharpest contrast anywhere in the roundup: asked about AI’s impact on U.S. society 20 years out, 56% of AI experts expect a positive result, versus just 17% of the general public. On one point the camps match, though — majorities on each side feel they have little or no control over the AI in their lives, and majorities on each side want more. That is a product lever: explainability and a visible off-ramp narrow the trust gap more effectively than raw capability does. Perceived voice in AI design splits the same way: 40% say White adults’ views are considered at least somewhat well, against 25% for Asian adults, 19% for Black adults, and 17% for Hispanic adults, while another 37% to 44% are unsure.
The Product-Team Takeaways
Start from the trust deficit, because it operates as a hard design constraint. Half of your users arrive already worried, so anything that widens AI’s reach — children’s products and emotional-support contexts above all — gets put under a microscope. Teams building for teens should treat the 12% emotional-support figure and the six-in-ten cheating perception as lines that must not be crossed. News features call for conservatism next: a mere 9% of adults get news from chatbots at least sometimes while 75% never do — and inside that small group, about half at least sometimes encounter inaccurate news there, with 16% seeing it often. Regulatory trust splits as well: 44% of Americans trust the U.S. government to regulate AI while 47% do not, yet the figures run 54% among Republican leaners against 36% among Democratic leaners — no single narrative covers the whole U.S. market. Age, finally, segments better than treating “users” as one mass: ChatGPT use at work among 18-to-29-year-olds (38%) runs at more than double the 50-plus rate (18%), so products skewing young can default to AI-native experiences while older audiences need plainer explanation, demonstration, and an exit.