Artificial Intelligence (AI) Polls of Americans: What the 2026 Surveys Found

Majorities of American adults are wary of artificial intelligence (AI), according to three national polls released in 2026. The polls asked different questions, so their percentages measure different things and should not be ranked against each other.

Pace, personal feeling and what companies should do are three separate questions, and each poll's figure answers only its own. Pew's trend runs from 2021 to 2026.

What the 2026 Polls Found

Each of the three 2026 national polls reached a majority on its own question. The AP-NORC poll found that 64% of adults believe AI technology in the United States is developing too quickly.

Quinnipiac University asked what AI companies in the United States should do about pace. Of its respondents, 77% chose stop or slow down: 30% said stop developing powerful AI systems until their safety can be evaluated, and 47% said slow down until safety can be evaluated.

Pew Research Center asked whether increased AI use in daily life makes people more concerned or more excited. 52% said more concerned, 37% said equally concerned and excited, and 9% said more excited than concerned.

The three figures measure pace, company action and personal feeling.

AP-NORC Findings on Pace, Party and Priorities

AP-NORC's pace question drew majorities in both parties. Democrats gave a higher share (76%) than Republicans (54%) that AI is developing too quickly.

The poll was fielded September 24 to 28, 2026. It interviewed 2,140 adults online and by telephone through the AmeriSpeak Panel, a probability-based panel of NORC at the University of Chicago. The overall margin of sampling error is plus or minus 2.9 percentage points.

Among all adults, 64% say AI technology in the United States is developing too quickly. Another 8% say it is developing too slowly, and about a quarter say the pace is about right. The source does not give the party breakdown for the too-slowly and about-right answers.

On what the government's approach to AI should prioritize, Democrats and Republicans gave similar shares on three items. Those items are keeping AI under human control, protecting workers in the United States, and promoting technological innovation. The parties diverged more on minimizing environmental harm and on competition with other countries.

AP-NORC's own summary is that Americans are wary of how fast AI is developing. It also reports that large majorities in both parties want the government to prioritize keeping AI under human control.

Quinnipiac Findings on Pace, Harm and Trust

Quinnipiac University's national poll of adults, released September 30, 2026, asked respondents what AI companies in the United States should do about the pace of AI development. Stop or slow down drew 77% combined. The poll was conducted in collaboration with the Quinnipiac School of Computing & Engineering and School of Business.

Four options were offered. 30% chose stop developing powerful AI systems until their safety can be evaluated. 47% chose slow down until safety can be evaluated. 14% chose keep developing at the current pace, and 5% chose move faster. The 30% and 47% answers add up to the 77% combined figure.

On survival, 73% are concerned that future AI systems could potentially threaten human survival. That is 38% very concerned and 35% somewhat concerned. The other 25% are not so concerned (14%) or not concerned at all (11%).

On impact, 53% think AI will do more harm than good in their day-to-day lives. 34% think it will do more good, and 13% offered no opinion. Quinnipiac describes that split as little changed from its March 30, 2026 AI poll.

On the source of harm, 52% are more concerned about AI being used by humans to do harm to humanity. 33% are more concerned about autonomous AI agents doing harm. 2% volunteered that AI would cause no harm, and 12% offered no opinion.

Quinnipiac also asked about trust in the leaders of AI companies: 74% have not so much trust (29%) or none at all (45%), and 24% have a lot of (3%) or some (21%) trust.

Pew's Five-Year Trend and Its Party Shift

Pew's share of adults more concerned than excited about AI rose from 37% in 2021 to 52% in 2026. The survey was fielded June 22 to 28, 2026 among adults in the United States, and Pew published it September 16, 2026.

The question asked: 'Overall, would you say the increased use of artificial intelligence (AI) in daily life makes you feel more concerned than excited, more excited than concerned, or equally concerned and excited?'

The series ran 37% in 2021, 38% in 2022, 52% in 2023, 51% in 2024, 50% in 2025 and 52% in 2026. In 2026, 37% said they were equally concerned and excited, and 9% said more excited than concerned.

Among Democrats, including independents who lean toward the party, 56% are more concerned than excited. Among Republicans, including leaners, 49% are. Pew says this is the first time Democrats are more likely than Republicans to express this concern.

Democrats rose and Republicans fell. Among Democrats, the share rose from 46% in 2023 to 56% in 2026, and it is up 25 percentage points since 2021. Among Republicans, the share fell 10 percentage points since 2023, to 49%.

Liberal Democrats rose from 45% in 2023 to 63%. Conservative Republicans dropped 14 percentage points over the same period. Pew says moderate and conservative Democrats and moderate and liberal Republicans have not changed significantly since 2023.

Pew reports that AI has emerged as a midterm election issue, with candidates in both parties debating data centers, government regulation of AI and AI's effect on jobs. Related reading on this site: the data center impact explainer.

Why the Percentages Differ

The percentages come from different questions with different answer options. The table lists the three questions side by side.

The polls were also fielded at different times. Pew's survey ran in June, and AP-NORC's ran in late September. Quinnipiac's field dates are not listed here.

Pew asked about feelings and AP-NORC asked about pace, so their numbers describe different things.

A poll records stated opinion at the time it was taken. The Pew and Quinnipiac questions record stated opinion, not behavior, and AP-NORC's questions on using AI for voting information are the only usage figures here.

Guardrails, China Competition and Party Trust

Quinnipiac's answers on AI safeguards and on competition with China were both majority answers in the same poll.

91% say it is very important (74%) or somewhat important (17%) that the United States establish guardrails for AI systems. 7% say it is not important. 86% support requiring AI developers to meet independent safety standards even if it slows development, and 9% oppose.

69% say it is very important (40%) or somewhat important (29%) that the United States keep up with China in developing AI systems. Both questions come from the same Quinnipiac poll. The same respondents can want both safeguards and competitiveness.

AP-NORC asked about trust in the parties. 23% trust the Democratic Party more than the Republican Party to do a better job handling AI. 21% trust the Republican Party more, and 39% say neither party can be trusted to do a good job.

The same poll found 31% approve of how President Trump is dealing with AI. On voting, 21% have used or plan to use AI to learn about a candidate or an issue on the ballot. Fewer say they have used or plan to use it to find how to vote (12%), how to register to vote (10%), or to decide who to vote for (10%).

Related coverage on this site: the AI backlash page and the AI legislation page.

Checks to Run on an AI Poll Headline

Four details show what a poll headline covers: the question, the sample, the margin of error and the field dates.

1. Question wording. Pace, company action and personal feeling each produced a different figure in these three polls. Confirm which one the headline reports before comparing numbers.
2. Sample and margin. AP-NORC's 2,140 adults carry a margin of sampling error of plus or minus 2.9 percentage points. The sample sizes and margins for the Quinnipiac and Pew polls are not listed here.
3. Dates. Pew fielded its survey June 22 to 28, 2026, and AP-NORC fielded its survey September 24 to 28, 2026. Quinnipiac released its poll September 30, 2026, and its field dates are not listed here. Figures from different months describe different points in time.
4. What a poll cannot show. These polls record stated opinion at the time they were taken. They show stated opinion at the time, not what people will do next, and this page has no data on opinion outside the United States.

Read any new AI poll against this list, starting with the question wording. The releases for the three polls above: the AP-NORC release, the Quinnipiac release and the Pew short read.

FAQ

Do most Americans think AI is moving too fast?

In the AP-NORC poll, 64% of adults said AI technology in the United States is developing too quickly. 8% said too slowly, and about a quarter said the pace is about right. The poll was fielded September 24 to 28, 2026, with a margin of sampling error of plus or minus 2.9 percentage points.

What did Pew find about concern over AI?

Pew Research Center found 52% of adults more concerned than excited about AI in daily life in its survey fielded June 22 to 28, 2026. That share was 37% in 2021 and 52% in 2023, and it was 52% again in 2026. In 2026, 37% said they were equally concerned and excited, and 9% said more excited than concerned.

What did the Quinnipiac AI poll find?

The Quinnipiac University national poll released September 30, 2026 found 77% said AI companies should stop or slow down powerful AI development until safety can be evaluated, with 30% choosing stop and 47% choosing slow down. It also found 86% support independent safety standards for AI companies, even if they slow development. 74% have not so much trust or no trust at all in the leaders of AI companies.

Are Democrats or Republicans more worried about AI?

In AP-NORC's pace question, 76% of Democrats and 54% of Republicans said AI is developing too quickly. In Pew's concern question, 56% of Democrats and 49% of Republicans said they were more concerned than excited, and Pew says this is the first time Democrats were more likely than Republicans to express this concern. These are different questions.

Why do AI poll numbers differ from one pollster to another?

Each pollster asked a different question. AP-NORC asked about the pace of AI development, Quinnipiac asked what AI companies should do about pace, and Pew asked whether AI in daily life makes people more concerned or more excited. Percentages from those questions measure different things and should not be ranked against each other.

Can AI poll results predict how Americans will vote?

The polls covered here do not say how people will vote. They record stated opinion at the time each survey was taken, not behavior or what will happen later.