AI is showing up in more Americans’ lives, but the public mood is heading in the opposite direction. A new Quinnipiac University poll finds 55% of Americans now say AI will do more harm than good.
That unease extends beyond daily life alone. The poll also found rising concern about the technology’s role in education, while its use continues to spread across research, writing, and workplace tasks.
Too much uncertainty, too little trust
Some uses of AI drew far more public skepticism than others.
In education, 64% of Americans said AI will do more harm than good, up from 54% last year, while health care produced one of the poll’s closest splits: 45% said AI will do more harm than good, compared with 43% who said it will do more good than harm.
Chetan Jaiswal, associate professor of computer science and associate chair of the Department of Computing at Quinnipiac University’s School of Computing and Engineering, said: “Americans are not rejecting AI outright, but they are sending a warning. Too much uncertainty, too little trust, too little regulation, and too much fear about jobs.”
AI use is growing faster than confidence in it
AI is becoming more routine in everyday use. The poll revealed that Americans reported using AI for a range of practical tasks:
- 51% have used it to research topics they are curious about
- 28% have used it to write something for them
- 27% have used it for school or work projects
- 27% have used it for data analysis
Only 27% said they have never used AI tools, down from 33% in April 2025.
Trust, though, is lagging well behind adoption. The poll found 76% of Americans trust AI-generated information only some of the time or hardly ever, while just 21% trust it most of the time or almost all of the time.
Concern over AI’s effect on jobs is climbing fast
Quinnipiac also found 70% of Americans think advances in AI are likely to reduce the number of job opportunities for people, with Gen Z being the most pessimistic group at 81%.
Respondents sounded more alarmed about the job market as a whole than about their own positions. Among employed adults, 30% said they are very or somewhat concerned AI could make their own job obsolete.
Tamilla Triantoro, associate professor of business analytics and information systems at Quinnipiac University’s School of Business, said people “seem more willing to predict a tougher market than to picture themselves on the losing end of that disruption.”
AI as a backup, not as the boss
There’s a clear limit to how much authority Americans are willing to hand over.
Even if an AI tool were proven more accurate than a human at reading medical scans, 81% said they would still prefer a combination of both AI and a human, while just 3% said they would rely on AI alone.
That same instinct carried into the workplace. Eighty percent of Americans would be unwilling to work in a job where their direct supervisor was an AI program that assigned tasks and set schedules.
For all the talk of faster adoption, the survey suggests many Americans still want AI kept on a short leash. They may use it, test it, and make room for it, but when the stakes turn personal, many still want a human hand on the wheel.
Newsom’s latest order shows a more cautious approach to how AI is used in California government deals.