HAMPDEN, Conn. — More Americans than ever are turning to artificial intelligence for research, data analysis, medical advice, and even companionship. But that doesn’t mean they trust it or view it positively.
Just over half of Americans have used AI to research topics of interest, an increase of 14 points from April 2025 (37%), according to a new Quinnipiac University poll. 27% of Americans report having used AI for school or work projects or data analysis, and an additional 24% report having used AI to generate images.
The ease with which AI can be used to create images may be one reason why Americans are deeply distrustful of this technology, with 76% of respondents saying they rarely or sometimes trust AI.
“The discrepancy between AI use and trust is striking. Fifty-one percent say they use AI for research, and many also use it for writing, work, and data analysis, but only 21% trust the information generated by AI most or almost all of the time. Americans are clearly adopting AI, but they are doing so with deep hesitation rather than deep trust,” said Chetan Jaiswal, associate dean of the university’s School of Computing.
Distrust of AI also means that a clear majority of Americans (80%) say they are very concerned or somewhat concerned about AI, while 18% say they are not very concerned or not at all concerned.
Millennials and baby boomers have the least trust in AI, with 39% of both groups saying they are very concerned. Slightly more baby boomers (43%) than millennials (42%) said they were somewhat concerned. The Silent Generation, those born between 1928 and 1945, were the least concerned about AI, although more than 70% of that group said they were still very or somewhat concerned.
More than half of Americans (55%) believe AI will cause more harm than good in daily life, and that number has increased to 64%, and they believe AI will cause more harm than good when it comes to education.
But when it comes to the topic of health care, Americans are divided, with 45% believing it will do more harm than good and 43% saying the opposite. But the numbers show that Americans overwhelmingly want human doctors involved in their medical care, with 81% saying they would trust a combination of humans and artificial intelligence, and 14% saying they would trust humans alone, even if AI could prove to be more accurate.
“This desire for a ‘second opinion’ from a human, even if it proves to be less accurate than AI, reflects the lack of trust in AI that we see across the poll,” said Brian O’Neill, associate dean of the School of Computing Engineering.
Overall, 70% of Americans believe that AI will reduce job opportunities, and that number increases the younger the respondents are. Among Gen Z (1997-2008), 81% said their employment would be affected, compared to 66% of baby boomers and 57% of the silent generation.
Additionally, 56% of Americans believe they can tell the difference between an AI-generated video or recording, and nearly 3 in 10 (28%) say they have shared a video they later learned was artificially generated.
The poll reflects responses from 1,397 adults surveyed from March 19 to March 23 using random number dialing from landlines and mobile phones and live interviews. The university said the margin of error was 3.3%.
