WASHINGTON, DC – Americans’ familiarity with artificial intelligence has grown over the past two years. Today, 70% say they know somewhat or very well about AI, up from 64% in 2024.
But instead of becoming more comfortable with the technology, Americans’ previously improved attitudes toward AI have lost momentum, according to the latest Business in Society study from Bentley University-Gallup.
More Americans than in previous years believe that AI is doing more harm than good and predict that AI will reduce the number of jobs in the U.S. over the next decade. They also have less confidence in companies to handle it responsibly.
Table of Contents
More and more Americans believe that AI is doing more harm than good
Although many Americans continue to recognize both the benefits and risks of AI to society at large, perceptions have become more negative this year, particularly among young adults.
The percentage of Americans who believe AI does more harm than good has increased from 31% in 2025 to 39% in 2026, nearly matching the 40% in 2023. At the same time, a slightly smaller majority of U.S. adults (52%) than in the past say AI is equally harmful and good, while the percentage who believe AI does more harm than good drops slightly decreased by 9%.
###Embeddable###
Familiarity with AI has increased across all age groups, as have negative perceptions of AI. But this negativity is particularly pronounced among younger adults. Nearly half of adults ages 18 to 29 (47%) now believe AI does more harm than good, up from 36% in 2025 and significantly higher than the 30% in 2023 and 27% in 2024. The 11 percentage point increase between 2025 and 2026 was the largest among all age groups.
These and other findings from the recent Bentley-Gallup poll suggest that young adults are becoming increasingly skeptical about AI. It also aligns with research from the Walton Family Foundation, GSV Ventures and Gallup that shows Generation Z adults have become more negative toward AI despite the adoption of the technology.
Trust in companies to use AI responsibly has fallen
Most Americans continue to express little or no confidence in companies’ use of AI, even as companies increasingly integrate the technology into products, services and workplaces. Confidence in companies’ use of AI has declined slightly this year, but is still slightly above 2023 and 2024 levels.
Meanwhile, 27 percent of Americans say they have “a lot” or “some” confidence in companies using AI responsibly, down slightly from 31 percent in 2025. While not a major decline, this has halted two years of gradual improvement.
###Embeddable###
The loss of trust between 2025 and 2026 will be particularly pronounced among younger adults. Among adults ages 18 to 29, the percentage expressing at least some confidence fell from 30% to 20%, while the percentage saying they had none at all rose from 29% to 41%. Republicans remain the most trusted political group, but also saw the largest decline from last year, with trust falling from 41% to 35%.
Concern that AI could cost US jobs is growing
Job displacement due to AI remains a major problem. Most Americans continue to expect that AI will lead to job losses and disrupt the labor market over the next decade. However, in 2026, these concerns will become even greater, as nearly eight in ten Americans (79%) say AI will reduce the number of jobs in the U.S. over the next decade, up from 73% in 2025 and 75% before that.
###Embeddable###
Concern has increased most among Americans at earlier and more advanced stages of their careers. Among adults ages 18 to 29, the share of those who believe AI will eliminate jobs has increased 13 percentage points over the past year, from 62% to 75%; Among adults ages 45 to 59, it rose nine points, from 75% to 84%.
Additionally, the percentage of adults ages 18 to 29 who believe AI will increase U.S. jobs in the next decade has fallen sharply, from 14% in 2025 to 5% in 2026, making their views more in line with those of other age groups.
Americans increasingly view AI as comparable to humans, but not better
Most Americans expect AI to reduce the number of jobs in the U.S. over the next decade and increasingly view AI as comparable to humans in a range of tasks ranging from driving a car to providing medical advice. However, Americans continue to believe that AI will perform worse than humans on every task they were asked about in 2023 and 2026.
###Embeddable###
Americans are less likely to say that AI performs worse than humans in multiple areas, while the opinion that AI performs better than humans has changed little. The biggest declines in percentages saying AI performs worse than humans are in driving (down seven points), financial advice (down six points) and medical advice (down five points). In each of these areas, more and more Americans have come to believe that AI is doing the job about as well as humans would like, rather than doing it better.
Providing hiring recommendations is the exception, as the view that AI performs worse than humans remains essentially unchanged. Helping students with homework or studying is the only area where the share of Americans who say AI performs better than humans increased slightly, from 26% to 29%.
Conclusion
After two years of improving attitudes toward AI, the latest Bentley-Gallup poll shows Americans have cooled toward the technology in 2026. The change is particularly evident among young adults, who have become more skeptical about several of the study’s metrics, including the overall impact of AI, companies’ use of AI and its impact on jobs.
When public trust is already low, even small shifts in public opinion like this can matter as companies continue to expand the use of AI in products, services, and workplaces.
Find out more about how the Bentley University-Gallup Business in Society Research papers.
Stay up to date with the latest insights by following @Gallup on X And on Instagram. Find out more about how the Gallup panel functions.
###Embeddable###
https://news.gallup.com/poll/712751/americans-cool-toward.aspx
