AI and data centers are incredibly unpopular in every poll

Public opposition to the infrastructure behind artificial intelligence is emerging as a politically awkward issue ahead of the US midterm elections. A New York Times/Siena University survey released Tuesday found that 61 percent of likely voters objected to building data centers for AI systems. The result adds fresh evidence that enthusiasm for AI’s commercial expansion is not being matched by enthusiasm for the large, power-hungry facilities needed to run it.
The survey questioned 1,503 likely voters in early September, before a recent public debate over slowing AI-safety efforts. Only 14 percent of those polled described themselves as strongly supportive of new AI data-center construction. The finding points to a broad concern about the physical footprint of the technology, even as companies present AI as a central engine of future economic growth and public policy becomes more closely tied to its rollout.

Views differed sharply when respondents were grouped by their 2024 presidential vote. People who backed Donald Trump were nearly split, with 49 percent in favor and 45 percent opposed. Those who voted for Kamala Harris were much more likely to resist new facilities, with 74 percent opposed. Among people who did not vote, 67 percent said they opposed their construction. The divisions show that the issue crosses party lines, while still carrying different intensity among voting blocs.
Many opponents did not necessarily favor shutting down all development. Of respondents against data centers, 56 percent said they would prefer limits on construction, while 38 percent backed an outright prohibition. That distinction matters for officials considering zoning rules, permitting standards, utility planning, and local oversight. The polling suggests a substantial group may accept constrained growth, even as another large segment wants to stop projects altogether.
Environmental and water-use concerns were the most frequently cited reason for opposition, named by 32 percent of opponents. Another 21 percent pointed to effects on nearby communities. Distrust of, or dislike for, AI itself accounted for 18 percent. Concerns about inadequate regulation and the economy followed at 10 percent and 9 percent. The responses indicate that resistance is not centered on one complaint: it combines worries about resource demand, neighborhood consequences, governance, and the technology’s wider social role.
The latest results fit a pattern seen in other national polling. An NBC News survey in March placed AI’s popularity below that of Immigration and Customs Enforcement. A Gallup poll in May found that seven in ten Americans opposed AI data centers being built in their own communities. Together, the surveys suggest that the political challenge extends beyond abstract unease about algorithms to the more concrete question of where new facilities should be located and who bears their costs.
Despite the strength of those views, AI and data centers ranked far below kitchen-table concerns when voters were asked what would guide their midterm choices. For most demographic groups, fewer than 1 percent selected the issue. Voters aged 18 to 29 were a notable exception, with 3 percent choosing it. Economic conditions, foreign policy, inflation, and opposition to Trump or Republicans were among the issues selected more often.
Trust in handling the issue also remains closely tied to partisan identity. Respondents from each major party largely chose their own side as more capable of addressing it. Among the 554 people who answered that question, 40 percent trusted Democrats, 42 percent trusted Republicans, and 18 percent either did not know or declined to answer. The narrow overall divide leaves neither party with a clear national advantage on an issue that has not yet become a leading electoral priority.
That combination may shape the next stage of the debate. Opposition to new facilities is widespread, but its low ranking among decisive voting issues could limit the pressure on candidates to make sweeping national promises. Local fights over water, environmental effects, community disruption, permits, and regulation may instead remain the most immediate arena. As AI companies seek more computing capacity, officials will face a difficult balance between pursuing investment and answering voters who are increasingly skeptical of the data centers that make that expansion possible.
Source: The Verge
Related articles

Xbox’s first chief operating officer is out after just a few months
Helen Chiang is leaving Microsoft after 21 years.

Private tools for PDFs, Gmail and LLM cost audits
Privacy-first utilities for PDF price lists, bulk attachment downloads and API-cost checks.
Explore tools →
OpenAI now trains its frontier AI models with no internet access
OpenAI built an “offline world” for top AI models.

100+ browser tools, one fast directory
Open free tools for PDF, image, video, developer, finance, AI and SEO tasks — no sign-up or install.
Open WebHut →Create AI images with your own API key
aixipi runs on desktop/web, uses your own model API balance, and avoids subscription lock-in.
Try aixipi →