Pew survey: American skepticism toward AI

The defining tension in Pew's 2026 survey is not rejection but contradiction. Half of U.S. adults now use AI chatbots, up from about a third in 2024, and roughly a quarter reach for one daily ^[1]. Yet only 16% expect AI's societal impact over the next two decades to be positive, while around 40% brace for it to be negative ^[3]. People are folding the technology into their routines while withholding their faith in it. Pew's Jeffrey Gottfried captured the paradox plainly: usage is not endorsement, and Americans 'may use it, but they're still highly skeptical of it and how it will impact our society' ^[4]. That gap matters because it breaks the usual assumption that adoption signals approval. Habit, here, is running ahead of belief, and the product winning the most daily attention, ChatGPT at 44% of adults, is the same one feeding the unease ^[4].

Beneath the headline numbers lies a deeper structural divide: the people building AI and the people living with it see almost opposite futures. Pew's Dr. Eileen Yam has detailed how the optimism gap holds across nearly every dimension. On jobs, 73% of AI experts are positive against just 23% of the public; on productivity, 74% versus 17%; on whether AI will personally benefit them, 76% of experts versus 24% of the public, with 43% of the public expecting personal harm. The earlier August 2024 study set the same pattern in the long view, with 56% of experts seeing a positive 20-year impact compared to only 17% of the public ^[5]. What makes this chasm notable is its durability and its breadth: it cuts across political affiliation, race, age, and gender, suggesting public skepticism is not a partisan artifact but a broad-based judgment that the industry's confidence has failed to earn.

If Americans distrust AI's trajectory, the survey shows they also distrust the institutions meant to steer it. Sixty-seven percent have little or no confidence the U.S. government can regulate AI effectively, and 59% lack confidence that companies will develop and use it responsibly ^[1]. Neither the public sector nor the private sector commands a majority's trust to manage the technology, leaving a referee-shaped hole. The anxiety is concrete, not abstract: roughly seven-in-ten Americans (71%) predict AI will make their personal information less secure ^[2]. On X, Andrew Ng framed the broader pattern as a chasm between the AI community's optimism and the public's distrust, citing both Pew and Edelman findings, while TIME noted the rare cross-partisan consensus, from MAGA loyalists to Democratic socialists, that AI is simply moving too fast. On Reddit, the most-upvoted reactions traced the distrust to Big Tech's track record, energy and water costs of data centers, and AI-generated technical debt, echoing the survey's institutional skepticism rather than a fear of the technology itself.

Conventional wisdom says young people embrace new tech and older people resist it; Pew's age data complicate that story. Adults 18-29 are the heaviest chatbot users, at 66%, versus 23% of those 65 and older ^[2]. But familiarity has not bred comfort: 48% of under-30s expect AI to negatively impact society, against 35% of those 65-plus, and 37% think it will hurt them personally ^[2]. One-in-five under 30 even turn to chatbots for emotional support, a level of intimacy that coexists with deep wariness ^[2]. The Reddit response amplified this thread, with one widely surfaced statistic noting only 14% of under-30s view AI positively. The generation most fluent in AI is, in this reading, the most clear-eyed about its costs, a dynamic that should worry anyone betting that adoption alone will convert skeptics into advocates.

The survey's force comes from how consistently the numbers point one direction. Just 16% of Americans expect a positive 20-year societal impact, against roughly 40% who expect a negative one ^[3]. About 63% say AI is advancing too quickly, while only 19% say the pace is right and a mere 2% say it is too slow ^[4]. On trust, 67% doubt the government's ability to regulate AI and 59% doubt companies will act responsibly ^[1]. On personal stakes, 71% expect AI to erode their data security ^[2]. Adoption tells the counter-story: about half of adults now use chatbots, up from a third in 2024, led by ChatGPT (44%), Gemini (24%), Copilot (17%), and Meta AI (14%) ^[4]. And the expert-public split remains the widest gap of all, with 56% of experts versus 17% of the public foreseeing a positive long-run impact in Pew's 2024 study ^[5].