A new joint survey by Bentley University and Gallup, which has tracked US public opinion on artificial intelligence since 2023, finds that Americans report greater familiarity with AI but growing skepticism. The data show a consistent pattern: the more people learn about and use AI, the less they tend to trust it.
Perceptions of harm versus benefit
According to the survey, 39 percent of respondents believe AI causes more harm than good, up from 31 percent in 2025. Among 18–29 year olds the share rises to 47 percent, a notable increase from 27 percent in 2024 and 36 percent in 2025. A majority (52 percent) say AI brings about roughly as much good as harm, while those who think AI does more good than harm have fallen to 9 percent.
Increasing belief in human‑level or superior AI performance
At the same time, more people believe AI can match or outperform humans in certain tasks. The survey highlights areas such as financial advice and helping students with homework as examples where public perception of AI performance has grown. Perceptions have also improved for AI’s capabilities in driving and offering medical advice, although human recommendations still command greater trust in those domains.
Seventy percent of respondents said they feel somewhat or very knowledgeable about AI, up from 64 percent in the prior survey.
Declining confidence in companies’ responsible use of AI
Public confidence that companies will use AI responsibly has eroded sharply: nearly three‑quarters of respondents said they “not very much” or “not at all” trust companies to do so. Distrust is especially pronounced among young people, where roughly four out of five share this view.
Politically, Republicans remain the most optimistic group, but their confidence is slipping as well. About 35 percent of self‑identified Republicans expressed some or significant trust that companies will use AI responsibly, down from 41 percent in 2025.
Job loss worries are widespread
Seventy‑nine percent of respondents believe AI will reduce the total number of jobs— the highest level of concern in the survey’s four‑year history. Concern is strongest among middle‑aged adults: 84 percent of those aged 45–59 think AI will decrease the number of jobs, up from 75 percent in 2025.
Implications
Proponents of AI had suggested that broader use and familiarity would foster greater acceptance. The Bentley University and Gallup data indicate the opposite trend: increased knowledge and exposure appear to coincide with greater skepticism and lower trust in both the technology and in companies that deploy it.
Methodology
This report summarizes the latest results from the Bentley University and Gallup public‑opinion survey series that has followed US attitudes toward AI since 2023.
(Source: Gizmodo summary of the Bentley University and Gallup survey.)
