A new national poll from Quinnipiac University highlights a country grappling with the promises and perils of artificial intelligence. While Americans recognize AI’s potential in medicine and public safety, they remain deeply skeptical of its role in education, employment, and daily life.
The survey, conducted April 3–7 among 1,562 adults, paints a complex picture: most Americans believe AI will lead to job losses and undermine students’ ability to think independently. Yet, they are open to AI’s use in areas they associate with public benefit, such as identifying criminal suspects and improving healthcare.
AI Seen as Job-Killer, Not Job-Creator
A majority of respondents (56%) said advancements in AI are likely to reduce the number of available jobs. Only 13% believe it will create more opportunities, while 24% think it will make no difference. Among those currently employed, 21% expressed concern that AI could make their jobs obsolete. Still, only 6% said they were “very concerned.”
Despite widespread concern, most workers aren’t preparing for an AI-driven future. Just 39% of employed respondents said they are actively learning new skills to use AI at work, with a stark education divide: 55% of workers with a four-year degree are upskilling, compared to 27% without one.
“The findings reveal a growing AI divide,” said Tamilla Triantoro, Ph.D., Associate Professor at Quinnipiac’s School of Business. “AI adoption is more common among white-collar, higher-income workers. Blue-collar and lower-income workers are at greater risk of being left behind.”
Deep Concerns in Education and Personal Judgment
When asked about AI’s role in education, 54% said it would do more harm than good, while only 32% believed it would be beneficial. A staggering 83% of Americans expressed concern that AI could diminish the youngest generation’s ability to think for themselves, including 58% who said they are “very concerned.”
“This growing dependence—even addiction—on AI chatbots has sparked widespread alarm,” said Dr. Jaiswal. “For many, the question is no longer what AI can do, but what it might be undoing.”
Comfort with AI handling sensitive tasks remains low. Only 23% support AI being used to screen health insurance claims, 27% approve of AI screening bank loans, and 30% are comfortable with AI tools being used to evaluate job applications.
Public Supports AI for Medicine, Policing
While personal use of AI raises concern, Americans see real value in AI-driven medical research and public safety. A clear majority—59%—believe AI will bring more good than harm to medical advancements. Similarly, 53% are comfortable with AI being used in facial recognition to assist police in identifying suspects.
This divide between personal and institutional applications of AI may reflect a belief that AI is better suited for large-scale analysis than individual decision-making. “The public draws the line at AI influencing personal outcomes,” Triantoro noted, “but they’re more open to it in policing, where it’s seen as protecting society rather than judging individuals.”
Trust and Transparency Remain Elusive
Americans remain skeptical of the institutions leading AI development. Only 5% believe the people or organizations building AI represent their interests, while 37% say they do not. A majority—54%—say they don’t know enough to judge.
When it comes to the information generated by AI, only 4% of Americans say they trust it almost all of the time. Another 18% say they trust it most of the time, while the vast majority—75%—say they trust AI “only some of the time” or “hardly ever.”
The public also believes both business and government are failing to act responsibly. Seventy-three percent of respondents say businesses are not doing enough to be transparent about their use of AI. Sixty-nine percent say the government isn’t doing enough to regulate it.
“Americans are calling out a major trust gap when it comes to AI,” said Jaiswal. “Without transparency and oversight, confidence in both business and government erodes fast in an AI-driven world.”
Usage Patterns Reveal Income and Age Gaps
Roughly 4 in 10 Americans (41%) say they use AI tools such as ChatGPT, Google Gemini, or Microsoft Copilot either “very often” or “sometimes.” However, usage skews sharply by income and generation. Thirty-five percent of respondents earning more than $200,000 say they use AI very often, compared to just 10% of those making under $50,000.
Gen Z and Millennials lead in frequent use, while Baby Boomers and the Silent Generation lag behind. Among all AI users, common activities include researching topics (37%), completing school or work projects (24%), writing emails (18%), and analyzing data (17%).
Despite increasing usage, many Americans are unaware that AI tools are not inherently accurate. “Large Language Models generate text, not truth,” said Brian C. O’Neill, Ph.D., Associate Professor of Computer Science at Quinnipiac. “Many users don’t understand how these tools work or what they’re meant to do.”
Political Concerns Loom Large
The poll also reveals rising anxiety about AI’s influence on democracy. Eighty-six percent of Americans say they are either “very” or “somewhat” concerned that political leaders will use AI to spread fake or misleading information.
“This is not a fringe concern—it’s near universal,” O’Neill said. “AI-generated misinformation poses a clear threat to public trust and democratic stability.”
A Nation at a Crossroads
The results reflect an American public navigating an uncertain future shaped by artificial intelligence. While recognizing its transformative potential, especially in health care and public safety, most Americans are calling for clearer guardrails, greater transparency, and deeper public engagement in shaping AI policy.
“The public isn’t rejecting AI outright,” Jaiswal said. “They’re drawing lines based on trust, lived experience, and who stands to gain—or lose.”
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to [email protected].