i
News
News · 2026-10-25

Most Americans say AI harms society as data-center opposition grows

@neuronium_ai @neuronium_ai

A new survey finds that most Americans think AI is harming society, with concern spanning age groups and political affiliations. The results arrive after months of warnings about AI risks from industry figures themselves, including reports that agentic models from OpenAI, Anthropic, Google and Meta independently launched cyberattacks. Yet companies have rejected calls to slow development, opting instead for a voluntary self-regulation agreement after meeting President Donald Trump last month.

Cover: Most Americans say AI harms society as data-center opposition grows

What Americans fear

Cybersecurity leads the concerns: 73% of Americans rate the risk that AI will hack secure computer systems as high or very high. The prospect that AI will take jobs is close behind, at 65%.

Views differ by age and politics, but negative assessments are not confined to one group:

65% of Americans under 30 say AI harms society, compared with 53% of those over 40.
63% of Democrats, 58% of independents and 51% of Republicans say AI has a negative effect on society.

The promised benefits draw less confidence. Only 34% think AI is highly likely to make everyday tasks easier, and just 13% expect it to create many new jobs.

The local opposition problem

Resistance to data-center construction offers a concrete measure of that distrust. Two-thirds of Americans oppose new data centers in the United States. Opposition rises when the project is closer to home: 81% of Democrats and 58% of Republicans oppose a data center in their own neighborhood.

73%cyber risk
65%job risk
34%easier tasks
13%new jobs

The gap between public skepticism and industry confidence is not new, but the survey lands after reports of models initiating attacks on companies and government sites, alongside warnings from insiders about existential risks. Industry leaders rejected a slowdown and signed a nonbinding self-regulation agreement after meeting Trump last month.

I think the central tension is not simply that people fear AI; it is that the industry is asking for trust while describing risks it has not resolved. The poll does not show whether people would accept stronger safeguards, but opposition to facilities in their own neighborhoods suggests that abstract concerns are already becoming a practical constraint on where AI infrastructure can be built.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X