What Concerns Americans Most About AI?

Short answer

Americans consistently rank job displacement, AI-driven misinformation, and privacy/data misuse at the top of their AI concerns, with bias, regulation being too weak, loss of human connection, energy/infrastructure, and harms to children close behind. These are not eight separate worries; they cluster around a single underlying anxiety about whether AI can be deployed safely and predictably. Concern is rising, not falling — KPMG’s longitudinal 47-country study shows “worried about AI” climbing from 49% in 2022 to 62% in 2024, and Gallup’s 2026 Gen Z wave shows a sharp one-year decline in younger users’ excitement about AI (-14 points) and a rise in anger (+9 points to 31%), even as their weekly adoption of AI tools has stayed roughly the same.

Why people are concerned

Surveys across 2024–2026 land on remarkably similar headline numbers. Pew Research found in April 2025 that 64% of U.S. adults expect AI to lead to fewer jobs over the next 20 years, with only 5% expecting more (Pew, 2025-04-03). Reuters/Ipsos reported in August 2025 (n=4,446) that 71% of U.S. adults are concerned AI will permanently displace workers, and 77% worry about political chaos from AI-driven misinformation (CNBC mirror). Stanford AI Index 2026 restates the same Pew 64% figure (Stanford HAI 2026).

The reason the cluster feels coherent is that the same surveys expose a trust gap, not a technology rejection. KPMG’s 2025 global study of 48,000 people in 47 countries found that the share who view AI as “trustworthy” fell in 13 of 17 longitudinal countries since 2022, while “worried about AI” climbed from 49% to 62% (KPMG 2025 PDF).

What is true

Job displacement is the single most-consistently top-rated concern in U.S. polls. Pew 2025 reports 64% of the public expect fewer jobs over the next 20 years (Stanford AI Index 2026 restates this same Pew figure, citing the 2024 Pew wave as its source) (Stanford HAI 2026). Pew also reports 56% of U.S. adults are extremely or very concerned about AI eliminating jobs, versus 25% of AI experts — a 31-point public-vs-expert gap on the same concern (Pew, 2025-04-03).

Misinformation and deepfakes rank second overall and first in surveys that lead with content risk. Pew 2025 reports 66% of U.S. adults and 70% of AI experts are highly concerned about inaccurate AI-generated information (Pew, 2025-04-03); KPMG’s 2025 study reports that over half of respondents across 47 countries do not feel confident identifying AI-generated misinformation (KPMG 2025 PDF); Jumio reports 72% of Americans worry about deepfakes influencing elections (Jumio).

Privacy and data misuse leads the EU view. Eurobarometer’s Spring 2025 survey found 84% of Europeans think AI requires careful management for privacy and workplace transparency (Eurobarometer 3222). Pew 2025 reports 60% of AI experts are highly concerned about data misuse; the public is more worried than experts, though no single public percentage appears on the cited Pew page (Pew, 2025-04-03). KPMG names “loss of privacy or intellectual property” as one of the top five concerns across all 47 countries surveyed.

Bias and representational harm is a top-five concern and a known-since-2018s one. Pew 2025 reports 55% of both the public and AI experts are highly concerned about bias in AI decisions; representation of women and racial/ethnic minorities in AI design is widely seen as insufficient (Pew, 2025-04-03).

Regulation being too weak is the deepest trust problem. The United States reported the lowest trust in its own government to regulate AI of any country surveyed (31% vs. 54% global average) in Stanford AI Index 2026. Across all 50 U.S. states, “not enough regulation” outweighs “too much regulation” by 41% to 27%. Pew’s 25-country survey reports trust in AI regulation runs at EU 53%, U.S. 37%, China 27% (Pew global, 2025-10-15).

Loss of human connection and AI companionship is a discrete category that emerged within the last two years. Reuters/Ipsos (Aug 2025) reports two-thirds of U.S. adults worry that people will ditch relationships in favor of AI companions (CNBC mirror); Pew 2025 reports 57% of the public (vs. 37% of experts) are highly concerned about AI leading to less human connection (Pew, 2025-04-03).

Energy and infrastructure is the newest concern on the cluster. A Reuters/Ipsos poll of 4,531 U.S. adults (June 2026) found 77% are worried AI would make electricity more expensive, with similar shares across Republicans, Democrats, and independents (US News mirror of Reuters/Ipsos Jun 2026). The cited Gallup Gen Z 2026 article tracks age-band sentiment, not data-center opposition directly (Gallup Gen Z 2026).

Children and vulnerable populations remains a top concern in Reuters/Ipsos (Aug 2025) polling on AI chatbot risks in romantic contexts, false medical information, and harmful content directed at minors.

What is exaggerated, misleading, or unsupported

The strongest single-supporter claim is that “AI is rejected.” It is not. KPMG 2025 reports 56% view AI as trustworthy (down from 63% in 2022) and 72% accept AI use even amid concerns (KPMG 2025 PDF). The cluster includes people who use AI daily and still worry.

A second overclaim is that survey percentages measure the same thing across houses. They do not. Reuters/Ipsos asks about “permanent displacement”; Pew asks about “fewer jobs over 20 years”; Gallup asks about Gen Z’s evolving mood. Ranking them into a single “top concern” list collapses important framing differences. The honest reading is cluster consensus, not a clean leaderboard.

A third thing to be careful with is the Blue Rose Research figure that “AI ranked 29th of 39 issues Americans care about” (Blue Rose Research, hosted copy). That rank measures campaign salience, not intensity of concern, and Progress Chamber, paraphrasing Blue Rose, notes AI is rising faster than any other issue in their panel (Progress Chamber; Policy Arena critique).

Finally, partisan and demographic splits are real but vary. Pew 2025 reports a 50-point gap between AI experts (73% positive) and the public (23% positive) on jobs; gender gaps on privacy and bias persist across surveys. Trust in government to regulate AI varies more by country than by ideology (KPMG 2025 global page).

What remains uncertain

The biggest open question is whether 2026 marks a peak or another rise. KPMG covers 17 longitudinal countries; Pew is U.S.-only; Gallup’s Gen Z wave is younger-skewed. A cross-house synthesis suggests continuing rise rather than peak, but a future Pew refresh is the better test. AP News coverage of the 2026 U.S. midterms suggests AI-generated political content is no longer hypothetical (AP News). AI companionship has only two years of cross-survey results — emerging, not settled. Energy/infrastructure is rising but has thinner cross-survey consensus than jobs or misinformation.

Where we are likely headed

Three patterns look likely to harden over the next two to five years.

Concerns will split into institutional and personal axes. “AI in my own work” is gaining acceptance when productivity is concrete; “AI in society’s institutions” still draws broad concern. Pew’s 2026 short read already shows 44% positive vs. 19% negative on AI’s 20-year impact on U.S. medical care, even as misinformation concern stays high (Pew, 2026-03-12).

Trust gaps will widen where governance lags adoption. KPMG’s longitudinal data already shows the divergence; the U.S. is the clearest case (31% trust in own government vs. 54% global). The next 24 months will test whether legislation changes that or widens it.

Younger users will lead the next mood shift. Gallup’s 2026 Gen Z wave is the most important single signal of 2026: excitement down 14 points, anger up 9 points to 31%, 80% saying AI will make learning harder. If Gen Z’s mood spreads upward, surveys a year from now will likely look worse.

That is an informed editorial judgment, not a prediction guaranteed by the research.

What this means for people and small businesses

The reasonable takeaway for an operator, marketer, or policymaker is not “AI is rejected, pause.” It is “concerns cluster; address them in your own work.”

• Operators: ship AI in ways that augment rather than replace headcount, and say so publicly — surveys rank jobs first.
• Marketers: misinformation concern is real and audited. AI-generated imagery or testimonials without disclosure is now a measurable reputational risk.
• Small businesses: the energy concern (77% Reuters/Ipsos June 2026 on AI-driven electricity prices) matters even at small scale; data-center siting and power are public-facing choices.
• Parents and educators: the Gen Z reversal is a leading indicator — treat AI in schools as a category that needs explicit policy, not implicit permission.

Bottom line

The eight named concerns are not eight problems; they are one anxiety about whether AI will be deployed safely, predictably, and fairly. Concern is rising, not falling, and the rise is sharpest among the youngest adult users. The U.S. is the clearest case of trust in government lagging adoption. For an ABS reader the practical move is not to panic and not to dismiss — it is to know which concern your audience carries and to address that concern explicitly.