Most Americans have made up their minds about AI data centers: keep them out of the neighborhood. A new NBC News Decision Desk survey finds 69 percent of respondents oppose building AI data centers in their own area, even as the technology keeps working its way into daily life.
The poll, conducted online among 7,105 adults from August 20 to September 1, found 70 percent of adults say they are more worried about AI than excited by it. About two-thirds say AI poses high risks to society, and 81 percent say the government isn’t doing enough to regulate it.
None of that skepticism has slowed usage. Fifty-two percent say they use AI tools very often or sometimes, up six points from a June 2025 survey. Those using it “very often” jumped to 23 percent from 14 percent a year earlier. Trust has not followed adoption: only 18 percent say they can trust AI-generated information most or almost all of the time.
Neither political party has convinced the public it can handle the issue. Forty-four percent trust neither party on AI policy, while 20 percent trust Democrats and 16 percent trust Republicans, 15 percent are unsure, and five percent trust both equally. Asked who should be responsible for protecting consumers from AI misinformation, 38 percent said government, 34 percent said the companies building the technology, and 28 percent said individuals themselves.
Jobs and Misinformation Top the Worry List
Misinformation is the biggest fear: nearly three in four Americans cite deepfakes, fraud and misinformation as their top concern, and two-thirds worry people are becoming too dependent on AI. One respondent, answering an open-ended survey question, wrote:
“It will become a Frankenstein beyond human control. It will dramatically widen the gap between the haves and the have nots.”
Jobs are another major flashpoint. Seventy percent say AI is taking away job opportunities, while just 7 percent say it’s adding them; 23 percent see no difference or aren’t sure. Among adults 18 to 29, 73 percent say AI is costing job opportunities. The worry breaks down along party lines, with 78 percent of Democrats and 69 percent of independents concerned about fewer job opportunities, compared with 62 percent of Republicans. One respondent put it this way:
“If we don’t heavily tax companies that use chatbots (and other AI uses) soon, we will eventually have companies without employees. It simply becomes a cost cutting exercise. That is probably not a good thing.”
Majorities See More Harm Than Good in Key Areas
Majorities of respondents believe AI will do more harm than good in K-12 education, news and information, elections and democracy, higher education, and national security. Pluralities expect more harm than help in the economy, healthcare and arts and entertainment.
Views split by subject and by gender. Fifty-five percent believe AI will do more good than harm in scientific research, versus 26 percent who see more harm than good, and 47 percent expect more good than harm in healthcare against 30 percent who expect more harm. Nearly two-thirds of men expect AI to benefit scientific research, compared with 45 percent of women, and men are twice as likely as women to think AI will help the economy more than hurt it.
Confidence in AI also depends heavily on the task at hand. Nearly three-quarters of respondents feel confident relying on AI-generated recipes or cooking instructions, but only about a third feel that way about AI-provided information on historical events or health conditions and medical treatments. Just 21 percent are comfortable relying on AI-generated content for breaking news or personal financial decisions.
