A pair of surveys released this week shows most Americans are using AI to save time on work and routine tasks, but still mistrust it and expect the technology to have negative impacts.
The Census Bureau reported that a March survey of 136,000 households found that 55% of workers reported using artificial intelligence on the job for at least one of 11 tasks. The top three were searching for information or technical help; writing communications, documentation or instructions; and generating ideas.
Roughly 86% of those using AI reported that it saved them at least an hour’s work on the job, while 13% said it didn’t save time or added to it.
The survey did not ask why people didn’t use AI — but it found that women, employees 50 or older and workers with only a high school diploma were less likely to use it than others.
“The Census Bureau plans to collect more data about AI in the coming months,” the federal agency said in an emailed statement.
The University of Pennsylvania’s Annenberg Public Policy Center reported that 1,320 adults it surveyed from June 16 to July 19 were split on the technology.
It found that 39% agreed with the statement that “AI’s impact on the U.S. will be negative,” 35% said it would be equally positive and negative, and only 18% said it would be positive. The remainder either did not know or didn’t answer the question.
Penn researchers found 61% of respondents were opposed to AI data centers, up from 49% in an earlier survey that ended in March. Most reported using the technology at least sometimes, but predicted negative impacts on jobs, children’s online safety, and personal privacy and security.
Medical research was the only area where most adults agreed that AI could help the nation.
Matt Levendusky, a Penn political science and communication professor who supervised the survey, said the findings confirm that people are still “working through the pluses and minuses” of the AI shift.
“People can use AI at work and think it saves them time but still think it’s a net negative for society,” Mr. Levendusky told The Washington Times. “These technologies have all emerged so quickly that none of us have fully unpacked their true costs and benefits.”
The two surveys released Tuesday are the latest to document a growing unease as chatbots reshape the economy and everyday life.
The National School Choice Awareness Foundation reported Monday that 76% of the 2,053 parents of school-aged children who responded to a July 24-31 online survey supported the use of AI in schools. But 48% said they also worried it could make students too dependent on technology.
“There is now another layer of figuring out what you can trust, what needs to be verified, and when AI is or is not appropriate to use,” Shelby Doyle, the advocacy group’s senior vice president of policy and national partnerships, said in an email. “Managers are also having to figure out expectations and boundaries in real time.”
Another survey of 1,006 workers that Resume Now conducted in June found that 74% trusted their own judgment over the responses of AI tools, which most reported using for basic research and planning.
An additional 12% told the career coaching firm that they didn’t trust AI at all, and 1 in 4 of those surveyed said they didn’t use it at work.
“AI is quickly becoming part of how work gets done, but these findings show workers do not want it to have the final say in workplace decisions,” said Keith Spencer, a Resume Now career expert.
‘Incremental gains’
Several industry experts not involved with the surveys said the findings confirm that most people have found generative AI difficult to learn and helpful only in limited situations.
“It’s incremental gains for the majority and more significant gains for those who have restructured the process of getting work done, not just adding more tools to the existing process,” said Darren Colclough, founder of Gameplan, a managed-marketing service for small and midsize businesses selling goods to each other.
Scott Foster, a Minnesota-based digital marketer and software creator, said the findings confirm that the technology learning curve has been steep.
“Education level makes a big difference,” Mr. Foster said. “People with bachelor’s degrees use it more and across more tasks.”
Officials at the National Artificial Intelligence Association, an industry trade group, urged AI users to be patient.
“What the numbers really reveal is that the divide is not so much generational as it is about exposure,” Caleb Max, the association’s president and CEO, said in an email. “The more time people spend using and experimenting with these tools, the more they feel comfortable using them.”
He pointed to a Goldman Sachs estimate that AI could lift global GDP by 7% over the coming decade — but only if adoption becomes widespread.
The Trump administration and officials in California, home to the nation’s most lucrative technology companies, have pushed for widespread AI adoption in workplaces and schools.
“The United States leads the world in AI innovation, and President Trump will keep it that way,” White House spokeswoman Liz Huston emailed Wednesday, commenting on the surveys. “The Trump Administration is doubling down on innovation to widen the gap between America and the rest of the world.”
Growing pains
Most early AI adopters say it could take years to sort out various ethical, privacy and practical issues around using generative chatbots.
Eric Turney, owner of the Oregon-based Monterey Co., said adopting AI has made tasks more difficult for his custom merchandise company’s sales, management, marketing and customer service roles.
“The roles getting harder are the ones where accuracy and trust matter,” Mr. Turney said. “AI can make a weak answer sound polished, and that is dangerous: One bad promise about pricing, artwork, production time or shipping can turn into a remake, refund or upset customer.”
Shama Hyder, who runs the executive advisory firm Rooster & Co., said most industries outside of healthcare are still “buying licenses” rather than embracing change.
“The complication is temporary,” said Ms. Hyder, a former global head of marketing for TikTok. “It’s what the middle of a transition looks like, and that’s exactly where we are right now.”
According to the experts, future surveys are likely to show gradual improvement in people’s productivity and comfort with using generative AI.
Sean Campbell, director of AI Integration for the MBA program at George Fox University in Oregon, said AI adoption will remain difficult only for those who resist it.
“Ultimately, we will all become more productive in AI use, either because we learn more about AI or AI learns more about us,” Mr. Campbell said. “Either way, we are going to end up in the same spot.”
Contact the author
Please read our comment policy before commenting.