Significant majorities of Americans oppose the construction of new data centers in the United States, believe that artificial intelligence will destroy many jobs and disapprove of President Donald Trump’s handling of the technology, according to a Washington Post-ABC News-Ipsos poll.
The findings released Wednesday show the gulf between a public with deep concerns about a technology reshaping American life and a president whose administration has embraced AI and data centers as a boon to U.S. economic vitality and national security that should not be bogged down by regulations. Majorities of both Republicans and Democrats said they had concerns about data centers and want more federal regulation of AI.
Less than two weeks before Election Day, the poll of more than 2,500 Americans also shows why AI and data centers have emerged as hot topics in many midterm races, including in the tight contests for Ohio governor and for U.S. Senate seats in Ohio and Michigan.
A 57 percent majority of U.S. adults said AI is a bad thing for society, with about one-third saying it is “extremely” or “very” bad. Two out of every 3 Americans oppose the construction of new U.S. data centers, the warehouses that house the powerful computer systems behind AI.
A majority of Republicans — and roughly two thirds of Americans overall — say the federal government is not doing enough to regulate the growth and development of AI. Some 67 percent say they disapprove of the way Trump is handling the technology, a similar pattern to public opinion of his policy on Iran, which has a 70 percent disapproval rating.
Numerous Republicans who are running in the midterms have distanced themselves from the White House’s approach to AI, particularly its championing of AI data centers.
Kyle Kondik, an elections analyst at the University of Virginia Center for Politics, said Trump’s usually strong political instincts are failing him on data centers. “I think the White House has a tin ear on this one,” he said.
White House spokeswoman Liz Huston said the Trump administration is helping turn AI and data centers into engines of prosperity with appropriate protections. “President Trump is committed to ensuring America continues to lead the world in super intelligence while protecting American consumers,” she said, using the administration’s preferred term for artificial intelligence.
Since the public debut of OpenAI’s ChatGPT four years ago, a growing number of Americans, schools, businesses and government officials have plunged into using AI. Meanwhile, the public is increasingly worried about the harms the technology may cause to people, jobs and the world, polls have found. (The Post has a content partnership with OpenAI.)
The fresh survey results show evidence of the contradictions in how Americans view the technology. While a minority of U.S. adults, 42 percent, believe AI is at least somewhat good for society, half say it’s at least somewhat good for them personally.
Mariah Arlt, a stay-at-home mother in Farmington, Missouri, is among those who acknowledge holding mixed opinions about AI’s drawbacks and benefits.
Arlt has found AI services to be imperfect but very helpful in a housing-related legal case, which she’d had difficulties navigating on her own, she said. But Arlt also said she has felt deceived by AI-generated videos that she had believed were real. She doesn’t want her two children to use AI as a substitute for human interactions, Arlt said.
Arlt, who is 33, said she wants legally mandated disclosures to ensure the public knows when material is made by AI. And she prefers AI confined to low-stakes uses, even as she uses it for her legal issue.
“When it comes to AI and computers, yes, they’re helpful. But they should more be for entertainment purposes,” Arlt said.
Even among the survey responses that reflected optimism about AI’s personal or societal benefits, the enthusiasm was relatively muted.
Of the people who said AI was good for them personally, 39 percent characterized it as “somewhat” good while a far smaller percentage said it was “extremely” or “very good.” Among those who said AI is bad for them personally, a combined 25 percent said it was “very” or “extremely” bad, compared with 24 percent who said it was “somewhat” bad.
When asked about the effects AI might have on society, less than one-third of U.S. adults said there was a “big” or “very big” chance that it would create many new jobs, make everyday tasks easier or help the U.S. achieve technology dominance. By contrast, more than 6 in 10 Americans said there’s a “high” or “very high” risk of AI eliminating many jobs. A similar majority said the same about the prospect of AI hacking secure computer systems and about the technology making people lose their ability to make decisions on their own.
About 1 in 3 Americans think there’s a high or very high risk of human extinction from AI, a fear of some AI industry insiders that garnered heavy publicity in recent weeks. For comparison, 37 percent of Americans are at least somewhat worried that they or someone in their family will be a victim of terrorism, Gallup has found.
There are signs that AI has become another issue that splits Americans along partisan and demographic lines, but the poll also found areas of agreement.
Republicans and men were more likely to say that AI is good for society and for them, compared with Democrats and women. About half of Republicans, versus 76 percent of Democrats, said they opposed the construction of new data centers in the U.S..
However, majorities of both Republicans and Democrats said they opposed new data centers in their communities. A majority of Republicans and Democrats supported more federal regulation of AI.
“To me it seems they aren’t doing anything, and that’s not acceptable,” Robby Strock, a 38-year-old pastor in Canton, Ohio, said about government inaction on AI regulation. Strock, who said he’s a registered Democrat, called tax breaks for data centers in Ohio a “huge mess up in state government.”
Nearly half of Americans say they don’t trust either party to handle AI. Democrats were trusted by 26 percent of respondents and Republicans by 19 percent.
Read detailed results of the Washington Post-ABC News-Ipsos poll. The poll was conducted online Oct. 9-13 among 2,545 U.S. adults nationwide reached through the Ipsos KnowledgePanel, an ongoing panel of U.S. households recruited by mail using random sampling methods. Overall results have a margin of error of plus or minus two percentage points. The sample was weighted to match population demographics, 2024 turnout/vote choice and political partisanship.
The post Most Americans give a thumbs down to AI and how Trump is handling it appeared first on Washington Post.
