AI is rapidly transforming the shape of businesses around the world and has explicitly been named by executives as a reason for layoffs. But most people don’t think it will come for their jobs.
According to a new poll released Thursday morning, many workers in several major economies predict that their own jobs and trades will be relatively unaffected by AI, even as they acknowledge that the technology will likely have large effects on society as a whole.
The survey, conducted by YouGov in partnership with Udemy, an online education platform, took the opinions of over 4,500 working-age adults across the United States, United Kingdom, India and Brazil.
While 70% of adults surveyed in the U.K. worry about economywide AI impacts, 39% worry about AI-related consequences for their own occupation. In the U.S., 72% of surveyed adults said they worried about the broader economic effects of AI, while 47% said they were concerned about their own jobs.
Elizabeth Weingarten, head of behavioral science insights at Udemy, said this observed spread in responses reveals psychological tendencies that have appeared in prior economic transitions.
“I think this shift is akin to a blind spot that many people had in the 1990s when the internet was first becoming more of a phenomenon,” Weingarten told NBC News. “There was this broad awareness that, ‘Yes, this is happening, and people are using it in certain corners,’ but not necessarily that this is going to disrupt every industry, and this is going to disrupt the way that I do my work.”
The YouGov/Udemy report also found that many workers feel they lack the AI skills necessary for an AI-focused economy.
This perceived skill gap was particularly apparent in the U.K., where 55% of workers said they had not received any AI training. At the same time, only 64% of British workers expressed motivation for career skill development.
“The AI train appears to be in motion,” Weingarten said. “It’s leaving the station, but people are not buying a ticket to ride.” Weingarten attributed the disparity to a phenomenon called the “awareness action gap.”
“People are saying, ‘I’m aware that this is coming, but I’m OK for right now. I can spend my spare time pursuing a hobby or thinking about saving money or whatever it is. I don’t have to take this action now or today.’”
Andrew Garin, an associate professor of economics at Carnegie Mellon University who was not involved in the survey, said that it broadly made sense for people to be unsure about how AI would affect their own jobs: “Even though we’ve seen the power of AI tools, the specific use cases in work where AI will be adopted successfully at scale is still unclear.”
“It’s normal to not want to invest money and significant time when you could be working and being productive, doing what you have to do right now, when you don’t know if AI skills are going to be valued,” Garin said.
Fourteen percent of adults surveyed in the United States and Brazil said they possessed adequate AI skills, and 47% of adults in the U.S. and 35% in Brazil said they had received no AI training at all.
Forty percent to 44% of workers in all surveyed countries said they had trouble learning to use specific AI tools, while 32%-40% of workers said they faced challenges incorporating AI into their existing workflows.
Garin said worker worries about retraining are also not unique to the rise of AI. “If we don’t know where the technology is going to be most transformative, it’s unclear what we should be training for.”
The new poll comes on the heels of several reports examining increased enterprise AI adoption.
A Deloitte survey of leading business executives from late October found rates of AI adoption were well above 80% — and increasing — throughout 2025, while a McKinsey report released last week found that 88% of respondents regularly use AI in at least one business function.
In a Pew Research Center report from February, 52% of surveyed workers said they were worried about AI’s potential role in the workplace.
For her part, Udemy’s Weingarten says the dueling narratives of persistent worker hesitancy about AI and AI skills despite increasing enterprise AI adoption fit into Thursday’s findings.
To handle the more daunting questions about AI transformation, Weingarten said that workers can take a pragmatic, piecemeal approach. Workers can “list the tasks that you do at work and research which specific AI tools can already perform a lot of those functions, and then get a more specific sense, grounding it in the real world, of how vulnerable you might be to disruption.”
Ultimately, Weingarten said, AI’s growing role in the workplace will require new approaches from both individuals and organizations. “Are we going to see organizations continue to step up and offer people training and make this technology more accessible?” Weingarten asked.
“I’m going to be looking both at individual human behavior and organizational behavior to see how things shift in the next couple of years,” she said.
