Most Americans believe artificial intelligence (AI) will do more harm than good in their daily lives, with concern rising sharply even as adoption accelerates and companies commit hundreds of billions of dollars to the technology.
According to a new Quinnipiac University poll, 55% of respondents expect AI to have a negative net impact on their day-to-day lives, up 11 percentage points from April 2025, while 70% believe advances in AI are likely to reduce job opportunities.
Surprisingly, only 7% expect the technology to create more jobs.
Widening gap
Overall, the findings, based on a mid-March survey of 1,397 adults, reveal a widening gap between rapid technological deployment and public confidence.
A notable souring of sentiment toward AI coincides with major technology firms ramping up their investment in the infrastructure on which AI operates.
Amazon, Meta, Google and Microsoft are collectively expected to spend around US$650 billion on AI-related infrastructure in 2026, positioning the sector as a central driver of U.S. economic growth.
Meantime, political spending tied to AI policy is also on the rise.
Venture capitalist Marc Andreessen and Greg Brockman have contributed tens of millions of dollars to U.S. midterm election campaigns backing candidates who favour limited regulation of AI, according to campaign finance disclosures and reporting compiled by OpenSecrets.
Employment and education
Drill down beyond the headline fear of AI, and public unease is especially pronounced around employment and education.
Almost two-thirds of respondents expect AI to worsen education outcomes, while only 27% said it would improve them.
Concerns about employment are also intensifying, with 70% of respondents now expecting AI to reduce job opportunities, up from 56% a year earlier.
According to analysis by job platform Indeed, entry-level job postings in the U.S. have declined by around 35% since 2023.
Consumer/corporate disconnect
However, there’s clearly a disconnect between deteriorating consumer sentiment toward AI and its continued rollout by corporate America.
While just under 27% of respondents claimed to have never used AI tools, more than half reported using AI for tasks like research, writing and data analysis.
However, trust remains low with 76% admitting to trusting AI either “rarely” or only “sometimes”, compared with 21% who claimed to trust it most or almost all of the time.
Chetan Jaiswal, a computer science professor at Quinnipiac, said the findings highlight a contradiction between adoption and confidence.
“Americans are clearly adopting AI, but they are doing so with deep hesitation, not deep trust,” he said in the university’s statement accompanying the poll.
Infrastructure flashpoint
Infrastructure development has also emerged as another flashpoint.
Sixty-five per cent of respondents oppose the construction of AI data centres in their communities, citing concerns over electricity costs, water consumption and noise.
This issue is expected to feature prominently in midterm election campaigns, which pose significant challenges for Republicans.
Warnings from industry leaders have reinforced public concern, with Dario Amodei, CEO of AI company Anthropic, recently claiming that AI could trigger an “unusually painful” disruption in the labour market, adding to anxieties reflected in the polling.
The survey also points to unease over governance and transparency.
Around two-thirds of respondents said businesses are not doing enough to be transparent about their use of AI, while a similar proportion said governments are failing to adequately regulate the technology.
AI oversight
The sentiment comes amid ongoing debate between U.S. federal and state authorities over the scope of AI oversight.
On national security, opinion is divided but tends to be negative.
A slight majority of respondents said they oppose the military using AI to select targets, while 36% support such applications.
The polling follows reports that the Pentagon has used AI tools in overseas operations, though details remain limited.
While generational differences are also evident, they do not fundamentally alter the overall trend.
Systemic risk versus personal vulnerability
Younger Americans report higher usage rates and greater familiarity with AI tools; however, they also express higher levels of concern about its economic impact.
Among respondents born after 1997, 81% said AI would reduce job opportunities, compared with 57% of those born before 1945.
Tamilla Triantoro, associate professor of business analytics at Quinnipiac, said the data suggests a disconnect between perceptions of systemic risk and personal vulnerability.
“Americans are more worried about what AI may do to the labour market than about what it may do to their own jobs,” she noted.
Overall, the findings indicate that while AI is becoming embedded in daily life, public acceptance is lagging behind its adoption.
Concerns over jobs, trust and regulation continue to shape the national debate as investment and political stakes continue to rise.
