More American workers are experimenting with artificial intelligence in their jobs, but skepticism is still widespread.

New Gallup polling finds that while more employees are using AI frequently in their work, there’s been an uptick in alarm that new technologies will replace their jobs. Many workers who are not using AI say they prefer to work without it, have ethical oppositions to the technology or worry about data privacy.

The poll, conducted in February, points to a divergence in how AI is reshaping American workplaces. Some find it to be a gamechanger for productivity and efficiency, while others are concerned about its potentially negative impacts.

Social worker Scott Segal said he regularly uses AI to find information that will help connect his elderly and vulnerable patients to health care resources in northern Virginia. While he knows that the human connection and care he brings to that work is important, he also believes that AI could soon replace him.

“I’m planning ahead,” said Segal, 53. “I think everyone who works in a replaceable field or trade should be planning ahead.”

Most workers using AI report productivity boosts

Roughly three in 10 employees are frequent users of AI in their jobs, meaning they use it daily or a few times a week. About two in 10 are infrequent users, using AI tools at work a few times a month or a few times a year.

The Gallup poll found that about four in 10 workers say their organization has adopted AI tools or technology to improve organizational practices. About two-thirds of those workers say AI has had an “extremely” or “somewhat” positive impact on their individual productivity and efficiency at work.

Workers using AI in management roles are more likely to say the technology has been at least “somewhat” positive for their productivity, compared with individual contributors. About seven in 10 leaders using AI at least a few times a year say AI has made them more efficient at work, compared with just over half of individual contributors.

Labor and employment attorney Elizabeth Bloch of Baton Rouge, Louisiana, said she uses ChatGPT to help “draft letters or emails in a diplomatic way because it’s a very adversarial profession and sometimes you get heated.”

AI tools appear to have a greater benefit for workers in managerial, health care and technology roles than in service jobs. About 6 in 10 employees in those fields who are using AI say it’s boosted their productivity at least “somewhat,” compared with 45% of those using it in service jobs.

Why some employees don’t use AI

Even when companies make AI tools available, there’s no guarantee employees will adopt them. About half of U.S. employees use AI only once a year or not at all, according to the Gallup study.

Bloch said she’s tried using AI for legal research but finds it is prone to hallucinations, or making up false information, even when using AI tools custom-built for legal work. She’s worried other lawyers who were already bad at finding and citing relevant case law are “going to be bad at using AI, because you’re not using the right prompts,” leading judges to sanction them for false citations.

Among workers who have AI tools available at their company and don’t use them, 46% say it’s because they prefer to keep doing their work the way they do it now. About 4 in 10 non-users who have AI available to them report that they are ethically opposed to AI, are concerned about data privacy or don’t believe AI can be helpful for the work they do.

About one-quarter of these non-users who have AI tools available say they have used AI at work and don’t find it helpful, while about 2 in 10 say they do not feel prepared to use AI effectively.

Thuy Pisone, a contract administrator in Maryland for a company that works with the federal government, said she uses AI weekly for mundane tasks but has avoided it for things she already can do just fine.

“I have heard from my colleagues that we could use AI to put together our PowerPoint slides,” Pisone said. “I’m a little biased in that, well, I could put my own PowerPoints together. I don’t need help because it took me time to hone up my skill.”

More workers are concerned about new technology taking jobs

While this was less of a reason for forgoing AI at work, the poll also found U.S. workers are increasingly concerned about being driven out of a job by new technologies.

About two in 10 — 18% — of U.S. workers say it is “very” or “somewhat” likely that their current job will be eliminated within the next five years because of new technology, automation, robots or AI. That’s up from 15% in 2025. People working at companies that have adopted AI are even more likely to be concerned that their job will be eliminated: 23% call this at least “somewhat” likely in the next few years.

A Fox News poll conducted in March found that about six in 10 registered voters believe AI will eliminate more jobs than it creates over the next five years. Only about one in 10 expect it will create more positions, and about one-third say it’s too soon to say. About seven in 10 employed voters say they are “not very” or “not at all” concerned their current job could be eliminated by AI.

Segal, the social worker in Virginia, said his alternative plan if AI replaces him is to start a new “health care chaperone service” that physically escorts patients from one appointment to another, especially when they’ve been sedated and don’t have family or others to pick them up.

“I don’t think that’s something that will be replaced for another maybe 10 or 15 years, until robots are embodied with AI,” Segal said. “I do believe that AI is going to displace most people’s employment functions and I question what people will do for livelihood at that point.”

In the meantime, he’s been asking AI chatbots to help him strategize on saving for his retirement.

Gallup’s quarterly workforce surveys were conducted with a random sample of adults age 18 and older who work full time and part time for organizations in the United States and are members of Gallup’s probability-based Gallup Panel. The most recent survey of 23,717 employed U.S. adults was conducted Feb. 4-19, 2026. The margin of sampling error for all respondents is plus or minus 0.9 percentage points.