According to the American Psychological Association, in spring 2025, nearly 47% of workers across all sectors reported using AI tools at least once a month to assist with their work, up from 34% a year earlier. work in america investigation.
“AI adoption isn’t just creeping up, it’s rapidly accelerating, and for nearly a quarter of employees, AI has gone from an experiment to something they do weekly,” said Dennis Stoll of APA’s Office of Applied Psychology.
About one in five workers feels pressured by their employer to use AI, and about three in 10 are worried they’ll fall behind if they don’t, Stoll said. “The feeling of having to deploy AI to keep up is a new form of stress in the workplace.”
Stoll and colleague Mark Chan discussed their findings at a recent National Academies webinar exploring how the U.S. can help workers learn new skills to adapt to AI-driven changes to work.
“Technology and automation are impacting the jobs available,” Margaret Baier, director of the Adult Skills and Knowledge Institute and dean of Rice University’s Department of Psychological Sciences, said at the event. “This will really require a focus on reskilling, which tends to be defined as acquiring skills for new roles, and upskilling to enhance skills for current roles.”
Human resource development is “chronically underfunded”
While the United States has no shortage of innovative training models or motivated workers and learners, it lacks a public funding system that can quickly create new training programs or scale up successful programs, explained Rachel Lipson, co-founder and scholar-in-residence of Harvard’s Workforce Project.
“Workforce development in the United States is chronically underfunded compared to its peers,” Lipson said. “Our active labor market policy spending ranks near the bottom at about 0.1% of GDP. This places us second from the bottom.” [Organization for Economic Cooperation and Development countries]It’s next to Mexico. ”
Lack of investment affects workers who are displaced by new technology and automation. “If you look at the last few waves of technological change and macro and structural change, it’s clear that the United States hasn’t done a particularly good job of supporting people in the transition from unemployment to new employment,” Lipson said.
“There’s a lot of research showing that we need a lot of social and psychological support to weather these winds of change,” she continued. Beyond its direct impact on income, unemployment can have long-term impacts on the health of workers, their families, children’s outcomes, and even the public safety of communities suffering significant structural job losses.
“There are many [of impacts] We’re hopeful that after the last few waves of change, we’ll move towards doing things a little bit differently this time around,” Lipson said.
Training tailored to “frontier” and “restructured” jobs
Looking ahead, Lipson explained that training will need to be tailored to the profession depending on which of three categories it falls into. The first category is “frontier” jobs, or completely new jobs created by new technology, in this case AI. Lipson noted that the term was coined by economist David Autor, whose research found that in 2018, about 60 percent of all U.S. jobs were in jobs that didn’t exist in 1940.
“There will be entirely new roles because the underlying technology itself is new,” she says. “We may need entirely new training programs, and we may need funding models that can support unproven programs, because these roles have not existed before.”
The second category is jobs that are “reorganized,” Lipson said, where the job title is the same but the skills within it are changing because new tools like AI are incorporated or the environment changes. “In some ways, I think this is going to be the most important thing for talent development,” she said. “In fact, what happens in the employer context will be very important. This includes apprenticeships and other upskilling models.”
The third category is legacy jobs. These are traditional occupations that will continue to be essential, such as tool and mold makers. While these jobs may be less affected by AI, Lipson said they still require attention as the workforce ages and retires. “We need to think differently to ensure we don’t lose training capabilities in areas that are still really important.”
Mr. Baier also pointed to the aging of the workforce, noting that workers 55 and older will be the fastest growing workforce for the foreseeable future, and that this demographic trend will impact the type of training provided for new skill jobs. Fluid reasoning ability and processing speed tend to decline with lifespan, she said. Research shows that older learners can be just as successful as younger learners when it comes to self-directed learning, but they often need to invest more time and effort to learn new skills.
Baier added that technology not only brings new challenges, but can also offer new ways of learning and training for individual workers through massive open online classes (MOOCs), YouTube videos, virtual and adaptive reality, and AI itself. “Machine learning really offers a lot of opportunities for adaptive and personalized learning,” she said.
Jaime Teevan, Chief Scientist and Technical Fellow at Microsoft, spoke about the organizational use of AI and emphasized the need to effectively incorporate AI into teams. Some early research on AI and teamwork suggests that not only can AI help improve individual performance, but that groups and pairs are more likely to collaborate with AI to produce better outcomes, she said.
“We need to intentionally design AI to consider collaboration as well as individual outcomes, and organizational leaders need to think about how they restructure their organizations to support that collaboration,” Teevan said.
Mr. Baier noted that reskilling efforts must also include workers who lack institutional support, such as those in the gig economy. “I think we need more support for unincorporated individuals to understand where they fit in the new economy,” she says. “What worries me is that we’re asking people to do so many things on their own and there aren’t a lot of resources available.”
Lipson said there could be more universality in the jobs affected by the AI wave, as opposed to previous waves of technological change, and that this could help create more public empathy for those dealing with unemployment due to structural change.
“If everyone recognizes the vulnerability of things that can happen beyond their control, it might help build some consensus on the types of support that can really help people navigate these changes,” she said.
Watch the webinar.
