Credit: Nathan Kuczmarski on Unsplash
In A Nutshell
- 65% of U.S. adults say the government has done too little to regulate AI, a view shared by majorities of Democrats, independents, and Republicans alike.
- Americans are most optimistic about AI in medicine, with 57% expecting a positive impact on medical research, but pessimistic about its effects on mental health, the economy, and creative arts.
- Nearly half of respondents, 49%, would oppose a new data center being built in their area, and 38% expect AI to drive up household utility costs.
- Most Americans aren’t panicking about losing their jobs to AI, but 28% say they are at least somewhat worried, and only 40% say they are not worried at all.
Artificial intelligence is everywhere right now, showing up in customer service chatbots, hospital software, hiring tools, and creative platforms, and Americans are paying attention. According to a new survey from the Annenberg Public Policy Center, roughly 78% of U.S. adults say they have heard at least a moderate amount about AI. But knowing about something and trusting that someone is managing it responsibly are two very different things. When asked whether the government has done too little, too much, or the right amount to regulate AI, 65% of respondents said too little. That’s not a fringe opinion. It’s a majority view shared across the political spectrum.
Conducted with 1,330 U.S. adults between February and March of 2026, the survey paints a portrait of a country that is cautiously watching AI’s rise, genuinely hopeful about some of its potential, and broadly worried that the rules aren’t keeping pace with the technology.
A Rare Moment of Bipartisan Frustration Over AI Regulation
Partisan agreement on almost anything is hard to come by these days, which makes this finding stand out. Among Democrats surveyed, 77% said the government has done too little to regulate AI. Among independents, that number was 72%. Republicans were less likely than Democrats or independents to say the government has done too little: 53% of GOP respondents held that view, while 39% said the government had done about the right amount and 7% said it had done too much.
When asked who should be driving that regulation, a slight majority, 52%, said the federal government should take the lead over state governments. A notable 33% of respondents said they weren’t sure, suggesting real public uncertainty about how AI governance should actually work. Republicans were more likely than Democrats to express that uncertainty, with 37% of GOP respondents selecting “not sure” compared to 28% of Democrats.
Americans Are Hopeful About AI in Medicine, Worried About Almost Everything Else
One of the clearest bright spots in the data is how Americans feel about AI’s potential in medicine. More than half of all respondents, 57%, said they expect AI to have a positive impact on medical research and discoveries over the next 10 years. That optimism held relatively steady regardless of political affiliation, making it one of the least polarized questions in the entire survey.
Things get murkier when the topic shifts to other areas. On AI’s impact on the overall U.S. economy, 19% expected a positive effect while 38% expected a negative one. Views on AI’s effect on mental health and well-being skewed pessimistic, with 48% predicting a negative impact compared to 17% who expected a positive one. Feelings about AI’s influence on movies, music, and other creative arts followed a similar pattern, 43% negative versus 22% positive.
U.S.-China relations drew the most uncertainty of all. Only 5% of respondents thought AI would have a positive effect on ties between the two countries, while 43% predicted a negative impact and 30% said they weren’t sure.
Most Americans Are Using AI, But Aren’t Panicking About Their Jobs
Despite all the headlines about automation threatening the workforce, outright panic about job loss is not the dominant mood in this survey. About 28% of respondents said they were either very worried or somewhat worried about losing their job or having their hours cut because of AI. A combined 40% said they were not too worried or not at all worried, and another 30% indicated they were not currently working, making the question less relevant to them personally.
Usage of AI tools is fairly widespread. About 38% of respondents said they used or interacted with AI at least almost every day in the past month, and only 16% said they had never used it at all.
Nobody Wants a Data Center in the Backyard
One concrete, local consequence of AI’s expansion is the construction of data centers, the massive facilities that power AI systems, and the public response was notably cold. Nearly half of all respondents, 49%, said they would oppose the construction of new data centers in their area, while only 21% said they would support it. Opposition was highest among Democrats and independents, at 57% and 54% respectively, while Republican opposition sat at 38%.
While the survey does not ask respondents why they oppose data centers, it does ask about one related concern: household utility costs. On that question, pessimism outweighed optimism, with 38% expecting AI to have a negative impact on utility bills versus 14% expecting a positive one.
For all the partisan noise that defines nearly every policy debate right now, this survey finds something unusual: broad, cross-party agreement that Washington is behind on AI. Whether that shared frustration eventually translates into actual legislation is a different question entirely. But if public opinion is any guide, the window for inaction is narrowing.
Paper Notes
Limitations
This survey relies on self-reported data, meaning responses reflect what participants say they believe or do rather than independently verified behavior. As a probability-based online panel, the survey may not fully capture the views of Americans who are not online or who are underrepresented in panel recruitment. The design effect of 1.68 indicates some added variability beyond a simple random sample, which is reflected in the margin of error of plus or minus 3.5 percentage points. Additionally, public opinion on a fast-moving topic like artificial intelligence can shift quickly, and these findings represent a snapshot from a specific window in early 2026.
Funding and Disclosures
The source document does not include specific information about funding sources or author disclosures beyond identifying the Annenberg Public Policy Center as the organization that conducted the survey. Readers seeking additional methodological detail are directed to the center’s full methodology report, referenced in the topline document.
Publication Details
Conducting organization: Annenberg Public Policy Center. Report title: Artificial Intelligence — Topline Report. Survey identifier: APPC W37 National Survey. Field dates: February 17 – March 20, 2026. Publication date: May 7, 2026. Sample size (unweighted): 1,330 U.S. adults. No individual authors, DOI, or peer-reviewed journal are listed in the source document.







