AIbaseSource publication

US Poll: 57% of Americans Believe AI Does More Harm Than Good, 70% Oppose Data Centers in Their Communities

A joint poll by ABC News, The Washington Post and Ipsos shows that most American adults hold a negative view of artificial intelligence.

A joint poll by ABC News, The Washington Post and Ipsos shows that most American adults hold a negative view of artificial intelligence. The survey, conducted October 9 to 13 with 2545 respondents (margin of error about ±2%), paints a rather bleak picture: 57% of respondents believe AI does more harm than good to society, while only 42% find it beneficial; 17% consider it "extremely harmful," while only 2% call it "extremely beneficial."

Attitudes divide clearly along lines of age and party. Among those under 30, 65% believe AI is harmful, falling to 53% among those over 40; among Democratic supporters this figure reaches 63%, while Republicans are relatively split, with 51% actually finding it beneficial. Intriguingly, when the question shifts from "society" to "yourself," the conclusion quietly reverses—50% believe AI is personally beneficial, while 49% find it harmful. In other words, many people worry about the macro-level impact while still relying on its practical value.

Data and Jobs Are the Biggest Concerns

Among specific risks, 73% worry AI will facilitate cyberattacks on security systems, 65% fear massive job losses, 62% worry humans will gradually lose the ability to make decisions independently, and 33% even view AI as a high risk of human extinction. By contrast, expectations of benefits are quite restrained: only 13% believe AI can create large numbers of new jobs, while 52% think this is nearly impossible.

This distrust also extends to bricks and mortar. 66% oppose building new data centers nationwide, and if one were built in their own community, opposition soars to 72%, reaching as high as 78% in rural areas. About two-thirds feel federal regulation is insufficient, 66% disapprove of Trump's handling of AI, and 49% trust neither party.

The core contradiction is plain to see: capital and computing power are pouring madly into AI, yet the public has not broadly bought in. The next test for companies is no longer just piling up chips and training models—it is proving that AI can genuinely benefit ordinary people, while containing the side effects on jobs, energy consumption, and community costs.


Original source

AIbase

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original