For the first time in twelve years of annual polling, artificial intelligence has entered the mainstream of American dread — not as an abstract workplace worry, but as a ranked, named threat to the political process and the broader economy. Chapman University's 12th annual Survey of American Fears, released Tuesday, placed AI-related concerns at No. 3 and No. 7 in the nation's top 10, two slots previously occupied by concerns that no longer made the cut, per Chapman's 2026 fear survey results.

The survey found that 51.9% of American adults fear AI could harm the political process, placing it at No. 3 overall — behind only corrupt government officials and the prospect of U.S. involvement in another world war. A separate question on AI's economic impact drew 49.7%, landing at No. 7. Both entries are new to the top 10 this year, and both arrived simultaneously — the first time AI has appeared in the top tier of the survey in any form.

"These findings make it clear that Americans are concerned not just about what AI might mean for their jobs, but about its potential impact on our political system and the economy," said Christopher Bader, Ph.D., a professor of sociology at Chapman University and the survey's lead researcher. "As artificial intelligence becomes more integrated into everyday life, understanding these fears will be increasingly important."

That concern is not unique to this survey. A June 2026 Pew Research Center poll found that 52% of U.S. adults say they feel more concerned than excited about the increased use of AI in daily life — up sharply from 37% in 2021. Only 9% say they are more excited, down from 18% five years ago.

America's Complete Fear Rankings in 2026

The survey, conducted online from April 7–21, 2026 by NORC at the University of Chicago using its AmeriSpeak® probability-based panel, asked 1,221 adults to rate their level of fear across dozens of items. The margin of error is 3.79 percentage points at the 95% confidence level. The full top 10, by share of respondents rating themselves "afraid" or "very afraid":

Fear of corrupt government officials held the 11th consecutive top position, a streak spanning multiple administrations and legislative cycles. What changed this year was the list below it: four of the top 10 entries are new, driven by a resurgent fear of global conflict, two new AI-specific questions, and the return of bioterrorism and physical attack concerns.

AI Enters the Apex of American Dread — and How the Questions Got There

The most significant contextual fact about the 2026 AI rankings is the one the top-line headline cannot convey: the questions themselves are new. The 2026 survey marks the first AI political questions asked about respondents' concerns regarding AI's effects on politics and the broader economy. Prior survey editions had asked about narrower AI concerns — in 2025, the relevant question was "AI replacing people in the workforce," which ranked 17th at 49.2% of respondents expressing fear.

That distinction matters for interpretation. The 2025 workforce question scored 49.2%; the 2026 political-process question scored 51.9% — a gap of about 2.7 percentage points within the margin of error. But the questions are not measuring the same thing. A reader who fears "AI replacing me at work" is expressing an individual economic concern. A reader who fears "AI harming the political process" is expressing concern about the integrity of shared democratic institutions. Survey researchers call this the difference between personal risk and societal risk — and societal risk perceptions tend to score higher because they are not tempered by optimism bias ("this bad thing will happen to others, not me").

What the 2026 results confirm is that when Chapman researchers gave Americans a question specifically about AI's systemic effects on democracy and the economy, more than half of the country said yes, that frightens them. The decision to ask that question at all reflects how far the public conversation about AI has traveled since Chapman last updated its question set.

Seth Benzell, Ph.D., an assistant professor at Chapman who studies the economics of AI and has co-authored policy briefs at Stanford's Digital Economy Lab on harnessing AI's productivity potential, offered a framing that captures the distinction: "People's fears about AI are much broader than concerns about automation or job displacement. The survey points to concerns about AI's effects on society, politics, and the economy, while also suggesting that greater familiarity with these tools may be associated with less fear," Benzell said in the release.

That last clause — "associated with" rather than "caused by" — is deliberate. The survey is cross-sectional rather than longitudinal: it captures a snapshot rather than a before-and-after. It cannot determine whether using AI makes people less afraid, or whether people who were already less fearful simply chose to adopt AI tools. Either reading has implications for how AI developers and policymakers approach the public trust problem.

Why AI Users Are Still Afraid

One of the survey's more telling findings involves the relationship between AI familiarity and AI anxiety. Among respondents who use AI tools at least weekly, 46% of weekly users alarmed expressed fear that the technology could harm the political process. Among those who have never used AI, 55% of non-users fear AI, with that figure rising to 55%.

The gap — nearly 10 percentage points — is real. But the ceiling is the more revealing number. Even among people who interact with AI regularly, almost half fear it could undermine democracy. Hands-on experience with consumer AI tools — chatbots, writing assistants, image generators — does not appear to fully resolve concerns about AI's broader effects on political information, electoral integrity, or economic distribution.

This is consistent with what risk perception researchers have long observed about societal-scale technological risks: familiarity with a technology's consumer applications does not eliminate fear of its systemic consequences. A person who uses a smartphone daily may still worry about surveillance capitalism. A person who uses a search engine may still worry about algorithmic manipulation of news. The same logic applies to AI: knowing how to use a chatbot does not tell you whether AI-generated political content is corrupting the information environment.

What Americans Hope For — and What Still Alarms

Beyond the top 10, the survey asked respondents whether specific technologies generate more hope or more alarm. Americans expressed three areas of net optimism: abundant clean energy, life-extension medicine, and human-robot collaboration in manufacturing. These findings suggest that technology anxiety is not monolithic — it is calibrated to perceived consequences, and consequences that feel distant or beneficial generate less fear than those that feel personal or politically destabilizing.

Three technologies generating more alarm than hope: continuous employee monitoring, genetically engineering human babies, and AI romantic companions. The first two have appeared in prior surveys and reflect enduring unease about workplace surveillance and bioethics. AI romantic companions represent a newer entrant to the conversation — and, researchers suggest, may be activating anxieties about social displacement and the erosion of human connection that existing survey frameworks were not designed to capture.

The full landscape of technology hope and fear visible in the 2026 data indicates that American public opinion is not anti-technology broadly — it is specifically alarmed about AI's potential to reshape political and economic power in ways that ordinary people cannot see or contest.

How the Survey Works — and Why the Results Can Be Trusted

The Chapman University Survey of American Fears is produced annually by the university's Wilkinson College research center survey, through the Ludie and David C. Henley Social Sciences Research Laboratory and the Earl Babbie Research Center. Chapman is classified by the Carnegie Classification system as an R2 high research activity institution and is based in Orange, California.

The survey's reliability rests substantially on how it is conducted: through NORC's AmeriSpeak® panel rather than a convenience or opt-in online sample. NORC at the University of Chicago recruits AmeriSpeak panelists using address-based sampling covering approximately 97% of U.S. households. This means every household has a known, nonzero probability of being selected — the defining criterion for any probability-based national sample. Surveys conducted on opt-in online panels, by contrast, capture only self-selected respondents and cannot be legitimately generalized to the population.

The 2026 survey used 1,221 completed interviews with a margin of error of ±3.79 percentage points.

Student researchers were involved in the analysis of findings, consistent with the survey program's educational mission.

Does Fear of AI Keep Rising? What the Broader Trend Shows

The Chapman survey data is a single-year snapshot, but it lands amid a documented multi-year trend of rising AI concern. Pew Research's June 2026 survey found that 52% of U.S. adults concerned more than excited about the increasing use of AI in daily life — up from 37% in 2021. The 2023 Pew survey had already registered a jump to 50% concerned, suggesting that the shift happened rapidly after the public launch of generative AI tools in late 2022 and has plateaued near majority concern since.

What the Chapman data adds to that picture is specificity: it is not just AI-in-general that Americans fear, but AI's effects on two particular systems they depend on — the political process and the economy. That specificity is significant for anyone designing AI governance frameworks or public communication strategies, because it points to where public trust needs to be rebuilt, not just whether trust is declining.

Frequently Asked Questions

What is the Chapman University Survey of American Fears?

Chapman University's Survey of American Fears is an annual probability-based poll now in its 12th year, conducted since 2014 by the university's Wilkinson College of Arts, Humanities, and Social Sciences. The survey asks a nationally representative sample of U.S. adults to rate their level of fear across dozens of items spanning government, technology, health, environmental, and personal concerns. The 2026 survey used NORC's AmeriSpeak® panel — a gold-standard probability-based sample — to interview 1,221 adults online in English between April 7–21, 2026, with a margin of error of ±3.79 percentage points.

Why did AI fears suddenly appear in the top 10 in 2026 when they weren't there before?

Two things happened simultaneously. First, the survey added new AI questions in 2026 that were not asked in prior years — specifically, questions about AI's effects on the political process and on the broader economy. The prior AI question (about AI replacing workers) ranked 17th at 49.2% in the 2025 survey. The new 2026 questions scored higher (51.9% and 49.7%) partly because societal-level fears tend to score higher than individual ones — people are often more willing to say "this could harm society" than "this will harm me personally." Second, the broader environment has shifted: generative AI tools are now embedded in search, productivity software, political advertising, and news production in ways that were not true when the prior question set was written. Both factors likely contributed.

What percentage of Americans are afraid of AI?

As of the 2026 Chapman survey, 51.9% fear AI harming politics and 49.7% fear AI could hurt the economy. A separate June 2026 Pew Research poll found 52% of adults more concerned than excited about AI broadly. Both surveys use probability-based national samples and can be legitimately generalized to the U.S. adult population. Using AI regularly makes a person somewhat less fearful — 46% of weekly users vs. 55% of non-users expressed fear of AI harming politics — but more than half of even regular users remain concerned about AI's societal effects.

Does using AI make you less afraid of it?

The Chapman data show a meaningful gap: weekly AI users 9 points less likely to fear AI's effects on the political process than people who have never used the technology (46% vs. 55%). But the relationship is correlation, not proven causation — the survey is cross-sectional and cannot determine whether AI use reduces fear or whether people who were already less fearful are simply more inclined to try AI tools. Researcher Seth Benzell specifically hedged: "greater familiarity with these tools may be associated with less fear." Either interpretation matters: if experience reduces fear, widening AI access could ease public anxiety over time; if less-fearful people are early adopters, the current familiarity gap may persist regardless of how many people eventually use AI, per Chapman's 2026 survey release.