Public Opinion on AI

Public-opinion research tracks how widely people use AI tools and how they assess the technology's effects. Through 2026, surveys in the United States record rising adoption of consumer chatbots alongside sentiment that tilts toward concern about AI's societal impact and the pace of its development. Instruments that repeat the same questions year to year add three further patterns: reported use and knowledge rising while trust in AI stays flat, economic and environmental expectations turning more negative, and majority support across parties for safety rules and independent safety testing. Outside the United States, Pew's 37-country survey of 2026 found that people in most countries surveyed expect AI to reduce rather than add jobs, with that expectation and awareness of AI both highest in wealthier countries (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)).

Survey instruments

The AP-NORC Center's press release and the Associated Press story give the field dates as July 6–24, 2026, while the AP-NORC project page gives July 7–24 (AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy; Source: apnorc.org).

Adoption

The Pew Research Center's "Americans and AI 2026" report, published June 17, 2026 and drawn from 5,119 US adults surveyed February 17–23, 2026, found that about half of US adults use AI chatbots, up from a third in 2024, with roughly a quarter using them daily. By product, ChatGPT led at 44% adoption, followed by Gemini (24%), Copilot (17%), Meta AI (14%), Grok (8%), and Claude (6%) (Americans and AI 2026: Chatbots, Smart Devices and Views on Impact).

The Gallup–SCSP study measures frequency of use rather than use by product. The share of adults using AI always, often or sometimes at work, at school or in their personal lives rose from 39% in its April–May 2025 wave to 51% in its June 2026 wave, and the share describing themselves as very or somewhat knowledgeable about AI rose from 60% to 68%. The growth came disproportionately from adults 30 and older: use rose 16 points among those aged 30–39, 13 points among those 40–64 and 16 points among those 65 and older, against 5 points among adults 18–29, who were already the heaviest users. Knowledge among adults 65 and older rose 14 points to 52%, against 80% among adults 18–29 (American AI Sentiment in 2026 (Gallup and SCSP)). In the 2025 wave, 98% of adults had seen or heard information about AI in the previous year and 8% described themselves as very knowledgeable (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). Gallup's quarterly workforce study, which covers employed adults only, recorded 50% using AI in their role at least a few times a year in the first quarter of 2026, 10 points more than in the second quarter of 2025 (American AI Sentiment in 2026 (Gallup and SCSP)).

Learning to use AI is mostly informal. In the 2026 Gallup–SCSP wave, 57% of adults had asked an AI for help with using AI, against 23% who had asked another person; 26% had used online training videos or articles, 16% had completed employer-provided training, 9% other courses or certification programs and 5% vocational or technical school programs (American AI Sentiment in 2026 (Gallup and SCSP)).

A later wave measured use by purpose rather than by product. Pew's survey of 3,488 U.S. adults conducted 22–28 June 2026 found 34% use chatbots for at least one of eight health and medical reasons — most commonly to get information quickly (28%), to work out what is causing symptoms (25%), and because the information is free or low cost (22%) (From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health). Health use follows the same demographic contours as overall chatbot use, reaching 56% among Asian adults, 44% among adults 18–29, 44% among four-year college graduates and 48% among upper-income adults, against 17% among adults 65 and older. It also tracks general disposition toward the technology closely: 60% of adults who say increased AI use makes them more excited are health users, against 21% of those who say it makes them more concerned (From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health).

Attitudes toward a specific application can run against personal use of it. In the same survey, pluralities said chatbots do more to hurt than help people using them for loneliness (39% to 19%), depression (36% to 17%) and stress (29% to 22%), with adults under 30 both the heaviest users of chatbots for emotional support and the most negative about the practice (From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health). See AI Mental Health and Psychological Harm.

Trust

The Gallup–SCSP study asks how much respondents trust AI's capability to make fair and unbiased decisions. In the 2025 wave, 2% trusted it fully and 29% somewhat, while 40% distrusted it somewhat and 20% fully (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). In the 2026 wave, 28% trusted AI fully or somewhat. The report describes that figure as "broadly in line with the 32% who said the same in 2025" and, in its conclusion, trust as "stable and low"; it does not describe the change as a decline. Sixty-five percent distrusted AI, and the share who fully trust it remained at 2% (American AI Sentiment in 2026 (Gallup and SCSP)). The 2025 report's component figures sum to 31%; neither report explains the one-point difference from the 32% baseline used in 2026.

Use is the strongest correlate of trust in both waves. In 2025, users were twice as likely as non-users to trust AI (46% against 23%), and the report's authors suggested that "greater adoption may build public confidence in AI technology over time" (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). In 2026, trust among adults using AI sometimes or more was 43% (46% in 2025), while trust among those using it rarely or never fell from 23% to 12%. The report attributes the flat total to the growing share of regular users offsetting the decline among infrequent users, and concludes that "higher use of and knowledge about AI tools are not associated with more positive sentiment toward it." Asked what would increase their trust, respondents gave similar support to greater transparency from companies about how AI makes decisions (49%), testing by independent experts (48%) and testing AI systems for fairness before public deployment (46%) (American AI Sentiment in 2026 (Gallup and SCSP)).

Sentiment and pace concerns

The same Pew study recorded views tilting negative: 40% of respondents expected AI to harm society over the next 20 years versus 16% who expected benefit, 63% said AI was advancing too quickly, and 71% believed it would make their personal information less secure (Americans and AI 2026: Chatbots, Smart Devices and Views on Impact). The privacy finding connects to AI and Privacy, and the pace concern to debates over Risk-Based AI Regulation.

A Quinnipiac University poll reported in July 2026 found 70% of U.S. adults believe AI will cost jobs and 55% believe it will do more harm than good in their daily lives (Source: wsj.com).

A Pew Research Center short read published August 18, 2026 found that 55% of U.S. adults under 30 were more concerned than excited about the increased use of AI in daily life, against 39% in 2024, the first time a majority of that age group had held that view; 11% were more excited than concerned. Across all U.S. adults the concerned-not-excited figure was 52%, up from 37% in 2021. Seventy-one percent of U.S. adults expected AI to lead to fewer jobs in the United States over the next 20 years, up from 64% in 2024; among adults under 30 the figure was 73%, up from 61%. The findings come from a survey of 3,488 U.S. adults conducted June 22-28, 2026, which included 569 respondents under 30, a subgroup carrying a margin of sampling error of plus or minus 4.7 percentage points (Sources: pewresearch.org; implicator.ai). Pew's 37-country report of September 2026, which draws on the same June wave, uses an 18–34 age band: 55% of Americans in that group were more concerned than excited, up from 40% in 2024, while the share was stable or lower among those 35–49 and 50 and older, leaving adults under 35 as likely as those 50 and older to be primarily concerned (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)). The employment expectations connect to AI Labor Disruption.

The AP-NORC Center and EPIC survey of July 2026, asking about the next ten years, found 52% of adults saying AI will do more to hurt than help society as a whole (44% in 2025) against 21% saying it will do more to help, 43% saying it will do more to hurt the U.S. economy (36% in 2025) against 22%, and 29% saying it will do more to hurt them personally (26% in 2025) against 22%. Democrats grew more pessimistic over the year: 57% of Democrats expected AI to do more to hurt society as a whole (46% in 2025) and 49% the economy (38% in 2025), against 50% and 40% of Republicans in 2026 (AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy).

A Reuters/Ipsos poll of 1,277 U.S. adults, concluded September 20, 2026 and published September 22, found 73% worried that AI companies have not gone far enough to prevent serious harm to society. Thirty-nine percent said AI is having a negative impact, up from 36% the previous month and the highest since the question was first asked in March 2026, against 11% who said it is having a positive impact. Fifty-five percent said slowing AI development would be good against 13% who said it would be bad, and 73% prioritized safe and responsible development against 23% who prioritized staying ahead globally; the margin of error is 3 percentage points (Source: reuters.com).

Economic expectations

The 2026 Gallup–SCSP wave recorded more adults expecting AI to reduce U.S. economic growth (45%) than to boost it (36%), which the report calls a reversal from 2025. Productivity was the one measure on which a majority (52%) remained optimistic. Adults aged 18–29 expected AI to reduce rather than boost growth by 57% to 26%, while adults 65 and older were the only age group in which more expected a boost (46%) than a reduction (33%) (American AI Sentiment in 2026 (Gallup and SCSP)). In 2025 the report had found optimism about productivity and growth concentrated among wealthier and more educated adults and pessimism about jobs and businesses concentrated among workers (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)).

Expectations for displaced workers are low. In 2026, 20% of adults expected that most or all workers whose jobs are replaced will be able to learn new skills in a different type of work (American AI Sentiment in 2026 (Gallup and SCSP)). The 2025 wave, asking about learning new skills "in a different industry," recorded 3% saying all such workers would, 20% most, 51% some, 24% a few and 2% none (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). The Gallup–SCSP job figure of 70% is close to the 70% in the Quinnipiac poll and the 71% in Pew's June 2026 wave, although the three questions are worded differently and cover horizons of ten years (Gallup–SCSP) and twenty years (Pew).

Annenberg two-wave measurement (2026)

The Annenberg Public Policy Center's W38 survey, fielded June 16 to July 19, 2026 among 1,320 U.S. adults (±3.48 points), is one of the few AI-attitude instruments published with a matched prior wave and a within-person change estimate, and its structure separates what moved from what did not. Opposition to new data centers in the respondent's own area rose from 48.9% to 60.6% over four months — +11.8 points wave to wave (95% CI +6.9 to +16.6), +12.5 among the 1,241 respondents interviewed in both waves — while three other repeated measures had confidence intervals spanning zero: government has done too little to regulate AI (65.4 → 67.9), too much (8.2 → 8.3), and AI's ten-year impact on the country will be negative (41.9 → 39.5) (Artificial Intelligence — Topline Report (APPC W38 National Survey)).

The reading the two halves support together is that a specific, locally sited grievance moved while the general dispositions around it stayed put. Data-center opposition also showed the lowest test–retest stability of the four repeated items (r = 0.55), consistent with an attitude still forming.

Its 13-area expected-impact table gives a ranking of where the public expects harm. Medical research and discoveries is the only net-positive area (+40.8). The most negative are personal privacy and data security (−62.8), children's safety online (−50.1) and employment and jobs (−46.5) (Artificial Intelligence — Topline Report (APPC W38 National Survey)). Heavy AI users are more positive than non-users on most areas — education and skill-development runs +13 net among heavy users against −35 among non-users — with privacy the exception, where both groups return 69% negative.

Environmental and data-center concerns

The AP-NORC Center for Public Affairs Research and the Energy Policy Institute at the University of Chicago (EPIC) released on September 16, 2026 a survey of 3,424 U.S. adults conducted in July 2026. It found 53% extremely or very concerned about the environmental impacts of AI, up from 41% in 2025, while concern about meat production, cryptocurrency and air travel was similar to 2025. Adults aged 18–29 moved from levels similar to other age groups in 2025 to the most concerned, rising from 38% to 60%. Forty-seven percent said AI will do more to hurt than help the environment, up from 33%. Democrats were more likely than Republicans to be extremely or very concerned about AI's environmental impact (65% against 42%), and Democratic concern rose from 49% in 2025 (AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy). See AI Environmental Impact.

The press release gives the shares at least somewhat concerned as 84% for local electricity prices, 79% for water supply and 78% for power outages; the factsheet's rounded components for water supply sum to 78%. Democrats were more likely than Republicans to be extremely or very concerned about electricity prices (70% against 58%), water supply (67% against 49%), power outages (61% against 46%) and noise (45% against 33%). Concern did not differ among urban, suburban and rural residents, although more data centers are planned in rural areas. About half of adults expected data centers to be at least somewhat beneficial for technological advancement (55%) and job creation (49%), and fewer for improved infrastructure (44%) and tax revenue (43%); Republicans were more positive than Democrats on technological advancement (60% against 54%), job creation (55% against 46%) and infrastructure (51% against 41%) (AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy).

Support for requiring developers to pay for grid upgrades was higher among adults 45 and older than among younger adults (80% against 66%) and among college graduates than among adults without a degree (83% against 68%) (AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy). The Associated Press reported the poll under the headline "There's one issue where Republicans and Democrats are unified before the midterms, a new poll finds" (Source: ap.org).

The energy question also divides by party in the Gallup–SCSP study, where Republicans were 16 points more supportive than Democrats of increasing U.S. energy production to build and run AI systems (American AI Sentiment in 2026 (Gallup and SCSP)). Local opposition to data-center construction, measured by the Annenberg survey above and by other polls, is covered at Data Center Siting / AI Power Politics.

Partisan polling has addressed the same question in campaign terms. A September 14, 2026 memorandum from Republican pollster Tony Fabrizio reports registered voters opposing more data-center construction 65% to 24%, and 61% preferring construction with conditions on power, ratepayer protection and water against 28% favoring a ban and 7% deferring to local officials. The memorandum does not name the survey's sponsor, field dates, sample size or question wording, and states that the three conditions were tested only as a package (Fabrizio Lee — Neutralizing Political Attacks on AI Data-Center Development Through Enforceable Conditions (memorandum, September 14, 2026)).

A Pew Research Center short read by Brian Kennedy, published September 22, 2026, states that "Americans have grown more negative toward the possible environmental and economic effects of data centers since the beginning of the year," and that a majority would be uncomfortable if a new data center started operating in their area; coverage of the finding of growing distrust among members of both parties appeared on September 23 (Americans' Views of Data Centers Have Turned More Negative (Pew Research Center, September 2026); Source: insideaipolicy.com). The survey of 10,548 U.S. adults on the American Trends Panel, conducted July 20 to August 9, 2026, repeated questions asked in January 2026. Four percent said data centers have a mostly good impact on each of the first three areas below, and between 21% and 24% said they are mostly good for local jobs and for local tax revenue, similar to the shares saying mostly bad. Between 18% and 32% were not sure of the impact in each area, and 12% said they had not heard of data centers, down from 25% in January; 35% had heard or read a lot about them, up from 25% (Americans' Views of Data Centers Have Turned More Negative (Pew Research Center, September 2026)).

The shift appeared across demographic and partisan groups. Among rural Americans, the share saying data centers are mostly bad for the environment rose from 32% to 50%, for home energy costs from 32% to 46% and for quality of life nearby from 28% to 45%; Pew notes that most data centers being planned or under development will be in rural areas. Among Republicans and Republican-leaning independents, 42% said data centers are mostly bad for the quality of life of people nearby, up from 24%, and Republicans were 7 percentage points more likely than in January to say the effects on local jobs and tax revenue are mostly bad, although more Republicans still called those effects mostly good than mostly bad. Among liberal Democrats, 34% said data centers are mostly bad for local tax revenue, up from 19%. Sixty percent said they would be not too or not at all comfortable with a new data center operating in their area, against 19% somewhat and 7% extremely or very comfortable; the uncomfortable share was 54% among Republicans, 68% among Democrats and 81% among liberal Democrats, with majorities in urban, suburban and rural areas (Americans' Views of Data Centers Have Turned More Negative (Pew Research Center, September 2026)). Negative views were more common among younger and more educated adults: 64% of adults 18–29 said data centers are mostly bad for the environment, against 42% of those 65 and older, and 66% of postgraduates against 41% of adults with a high school education or less. Among adults 18–29 the mostly-bad share rose from 23% to 37% for local jobs and from 19% to 35% for local tax revenue. Those who had heard a lot about data centers were more negative than those who had heard a little (73% against 54% on the environment), and 73% of the former would be uncomfortable with a new local data center (Americans' Views of Data Centers Have Turned More Negative (Pew Research Center, September 2026)).

Views on regulation, oversight and safety testing

Majorities in the Gallup–SCSP study put safety rules ahead of development speed. In 2025, 80% said the government should maintain rules for AI safety and data security even if AI capabilities develop more slowly, against 9% favoring speed and 11% unsure; support for maintaining rules was 88% among Democrats and 79% among Republicans and independents, and 97% agreed that AI safety and security should be subject to rules and regulations (Source: news.gallup.com). In 2026, 79% prioritized safety and data security rules against 7% favoring speed, a result the report describes as "unchanged from 2025" and holding across partisan lines (American AI Sentiment in 2026 (Gallup and SCSP)).

The 2025 figures are from Gallup's news release on the first wave (Source: news.gallup.com) and the 2026 figures from the second report, which does not itself compare them (American AI Sentiment in 2026 (Gallup and SCSP)). On the balance of influence over AI development, 42% said the government and businesses should have equal influence, 24% that the government should have more and 18% that businesses should (American AI Sentiment in 2026 (Gallup and SCSP)). The independent-testing results bear on the congressional debate over mandatory pre-release evaluation described at AI Pre-Release Vetting and Independent Verification Organizations (IVOs).

International comparisons

The largest cross-national measurement is the Pew Research Center's Spring 2026 Global Attitudes Survey, published September 17, 2026, which covered 42,151 adults in 36 countries interviewed February 8 to May 13, 2026 and combined them with the two U.S. American Trends Panel waves of February and June 2026 to report 37-country medians (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)).

In 34 of the 37 countries, people tend to believe AI will lead to fewer rather than more jobs in their country over the next 20 years; the medians are 46% fewer jobs, 13% not much difference, 9% more jobs and 25% unsure. Expectations of job loss are highest in Australia and South Korea (76% each), the United States (71%) and Greece (68%), and are correlated with GDP per capita: the median is 55% across 18 high-income countries against 36% across 18 middle-income countries, where more people are unsure (34% against 22%). The United States is the only country with an earlier reading, 64% in 2024. On inequality, the medians are 28% expecting AI to increase the gap between rich and poor in their country, 5% expecting a decrease and 36% unsure; no country has more than one in four expecting a decrease (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)).

Feelings about AI's increased use in daily life are mixed rather than predominantly negative: a 37% median are more concerned than excited, 41% equally concerned and excited and 13% more excited than concerned. Israel is the only country in which more people are primarily excited than primarily concerned (29% against 19%), and the shares primarily concerned are highest in Australia (53%), the United States (52%), Brazil (52%) and Canada (50%). Pew reports that concern and excitement are much less closely related to GDP per capita than expectations about jobs or inequality. The share primarily concerned rose between 2025 and 2026 in Sweden, Poland, the Netherlands, Hungary and Australia, and concern among young adults rose in six countries including the United States. Awareness is rising: a median of 31% had heard or read a lot about AI, with increases since 2025 in 11 of the 25 countries surveyed in both years (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)).

In many of the countries surveyed, people who have heard more about AI feel more optimistic about it; the comparison is not reported for the United States. The Gallup–SCSP study, which covers only the United States and measures use, knowledge and trust rather than awareness and excitement, concludes that higher use and knowledge "are not associated with more positive sentiment" (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026); American AI Sentiment in 2026 (Gallup and SCSP)).

In 12 mostly middle-income countries, Pew asked about trust in governments to regulate AI effectively. Median trust was 43% for China, 35% for the United States, 34% for the European Union and 56% for respondents' own country; China was the most trusted of the three in Bangladesh, Malaysia, Pakistan, Sri Lanka, and the West Bank and East Jerusalem, and the United States only in the Philippines. In Pew's 2025 survey of a different set of 25 countries, people had tended to trust the EU most, then the United States, then China (Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026)). See US-China AI Competition: Different Races, Different Metrics.

Views on the international AI race

A Pew Research Center survey of 3,488 US adults conducted June 22–28, 2026 and published July 23, 2026 found Americans saying by roughly three to one that China (36%) is more advanced than the United States (12%) in developing AI, with 18% calling them equal and 33% unsure. Forty-three percent said US AI leadership is extremely or very important, with wide splits by party (54% of Republicans against 34% of Democrats) and by age (56% of those 65 and older against 26% of adults aged 18 to 29). Fifty-one percent said AI will widen the gap between rich and poor countries, against 7% who said it will narrow (Source: pewresearch.org). See US-China AI Competition: Different Races, Different Metrics.

The 2025 Gallup–SCSP wave asked about competition with "other countries" without naming China. Seventy-nine percent said it is important for the United States to have more advanced AI technology than other countries, and 85% agreed that countries are currently competing for the most advanced AI. On current standing, 34% said the United States is keeping pace, 32% did not know, 22% said it is falling behind and 12% that it is moving ahead; 15% thought it very likely that the United States will have the world's most advanced AI in ten years. The share calling U.S. AI superiority very important rose with age, from 26% of adults 18–29 to 64% of those 65 and older. More favored developing AI with a large group of allies (42%) than a smaller group (19%) or alone (14%). Eighty-seven percent said foreign governments are likely to use AI to attack the United States within 20 years, and support for the U.S. military developing AI-enabled autonomous weapons rose from 39% to 53% when respondents were asked to assume other nations had developed them first (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). The 2026 report does not report the competition items (American AI Sentiment in 2026 (Gallup and SCSP)). See AI and National Security, Autonomous Weapons.

Views on platform oversight

Survey work on AI regulation sits alongside a broader appetite for platform oversight. A Reuters/Ipsos poll conducted online over the six days ending August 3, 2026 and published August 9, 2026 found that 61% of respondents said social media companies need firmer oversight, including 71% of Democrats and 62% of Republicans; support among independents was lower because many said they were unsure where they stood. Some 66% supported laws requiring social media companies to use age-verification tools to keep children under 16 off their platforms, with 74% of Republicans backing the idea against 69% of Democrats — one of the few measures on which Republican support ran ahead of Democratic. An overwhelming 85% said social media can be addictive for children, and a similar share said its use can harm their mental health. Respondents nonetheless reported liking the platforms: 70% said their time on them is fun against 35% who said it stressed them out, and they split 52% to 43% on whether they spend too much time on them daily. The poll surveyed 4,505 US adults with a margin of error of 2 percentage points in either direction (Source: reuters.com). See Age Verification, AI and Children.

The Gallup–SCSP study found the same direction of partisan difference on age verification: Republicans were more supportive than Democrats of mandatory age verification requirements in general, by 8 points, and of AI-based age verification on websites and apps, by 12 points (American AI Sentiment in 2026 (Gallup and SCSP)).

Reuters published the poll ahead of a trial expected to begin Wednesday, August 12, 2026 on state prosecutors' allegations that Meta designed its products to addict young users. Meta denies the allegations and has said it could face up to $1.4 trillion in penalties if the states prevail. The report does not name the case or the court, and notes that Congress has not passed legislation aimed at protecting children on social media, that more than a dozen states have passed laws regulating minors' social-media access, and that NetChoice — a trade group backed by Meta, TikTok and Snap — has sued to stop age-verification laws on free-speech grounds (Source: reuters.com). See NetChoice v. Bonta (CAADCA litigation), State of Tennessee v. Meta Platforms (Instagram design trial), New Mexico v. Meta Platforms (youth harm and chatbot decree).

Organized opposition

Beyond survey sentiment, an organized anti-AI movement has developed a hard-line wing, according to a July 11, 2026 Wall Street Journal feature. The feature centered on Oakland-based Stop AI, whose leader Sam Kirchner — an electrical engineering technician who led sit-ins at OpenAI — disappeared in November 2025 after telling a colleague "the ship may have sailed on nonviolence." The report described young professionals leaving careers for think tanks, nonprofits, and street protest against AI development (Source: wsj.com).

Partisan differences on regulation

Pew recorded a partisan reversal on confidence in the government's ability to regulate AI. Democrats' lack of confidence rose 20 points to 74%, while Republicans' fell from 70% to 61% (Americans and AI 2026: Chatbots, Smart Devices and Views on Impact). The shift tracks the period in which AI regulation became more closely identified with the federal executive, and bears on the preemption and state-law debates discussed in Techno-Federalism and State-Level AI Regulation.

On which party voters trust to handle the issue, a New York Times/Siena poll released on September 15, 2026 found Republicans ahead of Democrats by 42% to 40% on artificial intelligence, within the poll's margin of error. The Republican margin was +15 percentage points among voters aged 45–64 and −32 among those aged 18–29, and +14 among men against −6 among women. AI was the only issue in the poll on which Republicans held an advantage; Democrats led on the economy (+7), cost of living (+8), immigration (+7) and the deficit (+10) (Source: axios.com).

On specific proposals, the 2026 Gallup–SCSP wave found partisan differences whose direction depends on the proposal "rather than follow a single consistent ideological line." Beyond age verification and energy production, Republicans were more supportive than Democrats of expanding AI in manufacturing (by 12 points) and of reducing AI-related regulation (by 11 points), while Democrats were more supportive of increasing government funding for AI research at colleges and universities (54%, against 46% of independents and 44% of Republicans). Support for allowing companies or universities to use government resources for AI research was similar across parties (American AI Sentiment in 2026 (Gallup and SCSP)). SCSP's press release gives the university-funding figure for Democrats as 55% (Source: prnewswire.com).

Support for AI-related economic and workforce policy

The Center for Shared AI Prosperity published results on August 28, 2026 from a national survey of 56,000 respondents, conducted with Blue Rose Research, testing 79 economic policies responding to AI disruption. Net support is measured after respondents read short arguments for and against each policy and then choose support or opposition, with no undecided option. Of the 79 policies, 61 hold net approval after the opposition argument, the top dozen hold margins of at least 40 percentage points, and 48 hold 15 points or more; 31 of the 48 policies that mention AI had majority support, as did 19 of the 27 that mention AI economic disruption specifically. Labor-market and workforce-development policies drew significantly more support after argumentation than policies on taxation and revenue, income support and safety nets, or ownership and governance models. CSAIP states that it does not endorse policies by popularity, that "a popular policy is not the same as an effective policy," and that the survey cannot indicate which messages will persuade or how opinion will move (Source: blog.csaip.org). See AI Labor Disruption.

Workforce training is also the most supported AI policy in the Gallup–SCSP series. In 2025, 72% supported creating or expanding workforce training and education programs that teach how to develop or use AI, "equally popular regardless of political identification," against 56% for increasing federal spending on AI research, 55% for university AI research funding, 49% for letting companies or universities use government resources for AI research, 31% for tax incentives for AI developers and 23% for reducing rules on AI development (Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025)). In 2026, 72% supported promoting STEM education so workers are prepared to use AI in future jobs and 66% creating on-the-job AI training programs, with similar support across parties, while adults split 45% to 44% on teaching AI skills in core K-12 classes, with 27% strongly opposed against 14% strongly supportive (American AI Sentiment in 2026 (Gallup and SCSP)). See AI in Education.

Relationships

• supports: Americans' Views of Data Centers Have Turned More Negative (Pew Research Center, September 2026) — Pew's January–August 2026 tracking of views of data centers.
• related: AI and Privacy — data-security concern is a recurring driver of negative sentiment
• related: Risk-Based AI Regulation — pace concern informs regulatory-appetite debates
• related: Techno-Federalism — partisan split on regulation confidence
• related: AI Labor Disruption — workforce anxiety is a component of harm expectations
• related: Age Verification, AI and Children — the child-safety measures with the widest recorded support
• supports: From Diagnoses to Treatments, Why Americans Use AI Chatbots for Health — use-by-purpose figures and views on chatbot use for social and mental health support
• related: AI Mental Health and Psychological Harm, Healthcare — AI Deployment, Pew Research Center
• supports: Reward, Risk, and Regulation: American Attitudes Toward Artificial Intelligence (Gallup and SCSP, 2025), American AI Sentiment in 2026 (Gallup and SCSP) — the two waves of the annual Gallup–SCSP study: use, trust, economic expectations, regulation and testing preferences
• related: Gallup, Special Competitive Studies Project (SCSP), Annenberg Public Policy Center — producers of recurring survey instruments
• related: AI Pre-Release Vetting, Independent Verification Organizations (IVOs) — public preference for independent safety testing
• related: Data Center Siting / AI Power Politics, AI Environmental Impact — environmental and local data-center concern and support for developer-pays grid requirements
• supports: AP-NORC/EPIC 2026 Energy Survey: AI, Data Centers, Climate and Energy — the 2026 AP-NORC/EPIC survey: AI's environmental impact, expected effects of AI, and data-center attitudes by party
• supports: Globally, More People Expect AI to Cause Job Loss Than Growth (Pew Research Center, 2026) — 37-country survey: expected effects on jobs and inequality, concern and excitement, awareness, and trust in AI regulators
• related: AP-NORC Center for Public Affairs Research, Energy Policy Institute at the University of Chicago (EPIC) — producer and funder of the AP-NORC/EPIC survey
• related: US-China AI Competition: Different Races, Different Metrics, AI and National Security, Autonomous Weapons — views on the international race and military AI
• related: AI in Education — support for STEM, workplace training and K-12 AI instruction