The more Americans learn and interact with artificial intelligence, the more skeptical they grow of it. That is one trend that shows up in a new Bentley University-Gallup survey shared with Gizmodo. The examination of public opinion on AI, which dates back to 2023, found that while people increasingly see the technology as comparable to certain human capabilities, their trust in it and faith in its ability to improve their lives are declining. Per the survey, 39% of people surveyed believe AI does more harm than good, up from 31% in 2025. That figure climbs to 47% among 18- to 29-year-olds, marking a pretty significant uptick from the 27% who shared that negative view of AI in 2024 and 36% in 2025. The majority of those surveyed, 52%, believe the technology does equal amounts of good and harm, while the share who believe AI does more good than harm has fallen to just 9%. This comes as the general public becomes both more aware and, seemingly, more impressed with the performance of AI tools and agents. Seven in 10 people surveyed said they believe they are somewhat or extremely knowledgeable about AI, up from 64% last year. The survey also shows that the public increasingly believe that AI performs as well as or better than people in several areas, including providing financial advice and assisting students with homework. AI has also made significant gains in how people perceive its ability to drive a car and provide medical advice, though human recommendations still reign there.
Survey Finds Americans See AI as Comparable to Humans. It's Making Them Trust It Less
The better it gets, the worse it feels.
39% view AI as harmful (up from 31%), with 72% distrusting corporate AI—skepticism rising as awareness hits 70%. 79% fear job destruction (4-year high); tech leaders will face talent churn and deployment pushback as sentiment hardens.






