No one has a perfect roadmap to the future, but researchers at GovAI, a technology policy firm, and the Brookings Institution, a Washington think tank, used a new approach to estimate which workers are most likely to adapt to AI and which are least likely to adapt.
They concluded that many of the people most at risk if AI transforms work are also in the best position to find new jobs. You can use the search box and interactive graph above to find out which professions have promise and which don’t.
However, as history has shown, economists and researchers We’re bad at predicting how new technology will impact jobs and employees, so take these predictions seriously, not literally. Researchers are also ramping up research into AI in the workplace. Note that they are making a best guess that is informative but may be wrong.
Jed Kolko, a senior fellow at the Peterson Institute for International Economics, recently concluded that “all important questions about the impact of AI on the labor market remain unanswered.” Economists at Anthropic, the AI startup that developed the Claude chatbot, emphasized the need for “humility” in analyzing AI permeating the profession. (Humility is not common in Silicon Valley.)
As you scrutinize reports on AI-related job departures, you’ll be struck by glaring contradictions.
An influential Stanford University analysis last year found that AI is likely to take jobs away from young people in jobs such as software development and customer service. The adoption of AI was the fastest. Another study conducted by think tank Economic Innovation Group found essentially the opposite conclusion: in fields that have not been exposed to AI, such as fitness training and roofing, young workers in these occupations are outperforming other workers.
Although the Dallas Fed says AI is unlikely to displace jobs over the next decade, prominent CEOs continue to predict that AI will soon displace millions of jobs.
A flood of sometimes contradictory analyzes points to a yawning gap between how little is known about how AI is changing jobs And we all understand the need for certainty. This division allows Americans, business leaders, and policymakers to cherry-pick their preferred narratives. If you’re afraid of being thrown out for AI, there’s informed and uninformed evidence to fuel your nightmares. If you feel like your job is secure, there’s plenty of support out there.
There are two things that are generally agreed upon. Economists say there is currently no measurable evidence that Americans as a whole are losing jobs to AI. And while the victims of workplace automation in the past were primarily factory and trade workers, today it is white-collar jobs that are first on the list for AI reform.
A recent study led by GovAI senior researcher Sam Manning and colleague Thomas Aguirre took a novel approach.
They started with industry-standard measurements of AI “exposure” for more than 350 occupations. This is an estimate of how many work-related tasks, such as teachers grading homework, can be performed more efficiently by workers using AI.
These estimates found that skills used in computer programming, marketing, financial analysis, and customer service have a lot of overlap with AI capabilities, and therefore could theoretically make it easier for workers to be replaced by machines. (There is intense debate in economics about measuring AI “exposure” and whether high AI exposure helps or harms workers.)
The researchers went a step further by trying to quantify how easily people in jobs could move on to other, higher-paying jobs if AI were to take them away. The researchers reasoned that people with higher education and a variety of work histories would be able to change occupations more easily, and that being under 55 and living in cities where jobs are plentiful would help them become more affluent.
Web designers and secretaries both scored highly on the AI exposure survey, but their estimates of their ability to adapt were different. Manning and Aguirre found that secretaries are among the 6.1 million primarily administrative and administrative workers with the highest exposure to AI and the lowest estimated adaptability. (Manning is also a member of the Technology Policy Research Group at the National Innovation Foundation.)
The findings suggest that a large proportion of workers whose jobs might be changed or lost due to AI could be able to get back on their feet. However, fewer workers may have difficulty finding new jobs.
Researchers said women make up around 86% of the most vulnerable workers, suggesting that the negative effects of automation will not be felt equally across society.
Mark Muro, a senior fellow at Brookings University who assessed the study’s policy relevance, said the most vulnerable workers “may be out of sight and out of mind” for policymakers and the American public. The researchers cautioned that it is difficult to accurately assess people’s chances of finding other jobs.
Work mainly done by women
by men
Source: GovAI and Brookings Institution
Alison Elias, a professor at the University of Virginia School of Business, said changes in technology so far show why office workers, who are dominated by women, may be the losers of the AI revolution.
In her historical research, secretaries and other administrative staff often expected that new technology would free them up to perform higher-level tasks and help them advance their careers. Instead, workers are often expected to do more work for the same or lower pay, she said. They continued to report low job satisfaction.
“These people are in a very vulnerable position because they don’t have a lot of decision-making power over how the AI is used and the likelihood of exit is pretty low,” said Elias, who was not involved in the GovAI-Brookings analysis.
Economists say it’s nearly impossible to predict the impact of AI on the labor market based on the technology’s current capabilities or the business sectors in which AI will first penetrate. And they point to the track record of past technological revolutions, such as electricity and smartphones, that while eliminating some types of jobs, they have also created new jobs and economic growth that most people did not foresee.
This prediction was largely a departure from a prominent study more than a decade ago that estimated that nearly half of jobs could be destroyed by computer automation. Predictions that ATMs would wipe out bank tellers, early AI would wipe out radiologists, and player pianos would take away pianists’ jobs were wrong. Few people imagined that smartphones would bring new jobs in social media marketing and social media influence. And it’s unlikely that anyone is experiencing the 15-hour work week that economist John Maynard Keynes predicted in 1930.
“We don’t have a good track record of predicting how technological changes will play out in the labor market,” said Martha Gimbel, executive director of the Yale Budget Lab. It would have been hard to predict that the invention of electricity would lead to new jobs as elevator operators, and that a subsequent technological innovation, the button, would wipe out those jobs.
Another extinct profession, telephone operator, gives reason for both hope and pessimism about the impact of AI. Once one of the most common jobs for American women, the job disappeared as telephones modernized starting in the early 20th century, according to a 2024 research paper by James Feigenbaum and Daniel Gross.
The study found that switchboard operators who lost their jobs were much more likely than other switchboard operators to either be unable to find another job or to take a low-paying job. But within a few years, a boom in secretarial and restaurant jobs opened up new opportunities for young women. “I read it with some hope,” Feigenbaum, an economic historian at Boston University, said in an interview.
Feigenbaum doesn’t buy the argument that AI is significantly different for American workers than previous technological revolutions. The invention of electricity, the internal combustion engine and the internet were transformative technologies, he said, and “they’re not going to eliminate all jobs.”
Taylor Telford contributed to this report.
