Artificial intelligence is disrupting employment so much that many employees around the world are fearful of being laid off. A study by Israel’s Taub Center for Social Policy Research took an objective look at how AI is reshaping the composition of Israel’s unemployed population.
The researchers (Michael Debowie, Professor Gil Epstein and Professor Avi Weiss) found that while the impact of AI on overall unemployment remains limited, it is already changing who becomes unemployed. AI explains some of the changes in the occupational distribution of unemployed people between 2022 and 2025, especially after late 2024.
They found that the effects were concentrated in occupations that previously had very strong demand, low layoff rates, and persistent vacancies. Those occupations that had particularly low unemployment rates in 2022 are now experiencing the most significant increases in relative unemployment rates.
Increase in high-tech unemployment
“The era of impunity for high-tech workers is over. Our data shows that AI is tearing up the cards, explaining about a fifth of the increase in unemployment among programmers,” said Epstein, director of the Taub Center’s labor market policy program. He explained that the door is mainly closed to young people.While veteran staff become more efficient with the help of machines, it is the young people who will pay the price first.Those who wait for change and do not rush to improve their skills now will simply be left behind. At the back. ”
Ironically, lower-status blue-collar workers, such as barbers, garbage collectors, plumbers, house painters, and firefighters, and those whose jobs involve direct interaction with others, such as midwives, landscape artists, emergency medical technicians, and acupuncturists, will likely not be affected.
But it can have a big impact on bookkeepers. Lawyers and their paralegals and legal assistants. Junior market research analyst. clerks and administrators; Cashier, salesperson, product demonstrator, promoter. Warehouse and factory workers. This includes fast food and restaurant workers, as well as non-professional doctors, researchers, and computer scientists. There are drivers who have lost their jobs to self-driving cars, public relations professionals, and even actors and actresses whose jobs could be made redundant by virtual characters in movies and commercials.
Weiss, director of the Taub Center, an independent, nonpartisan research institute that deals with economic and social issues, added, “What we’re seeing here is a process in which technology is not only displacing workers, but completely changing the rules of the game. What that means for the unemployed is that there will be even more competition for existing jobs.” At the policy level, states must already launch support schemes for newly unemployed workers and workers by designing programs to provide them with skills that complement artificial intelligence so they can reintegrate into the changing labor market. ”
The findings show that the proportion of unemployed people in jobs at high risk of being ousted by AI has increased significantly. Between 2019 and 2022, these workers accounted for 14% to 16% of all unemployed people in Israel, but by 2025 that proportion rose to 20% to 25%. Vacancy rates for these positions are decreasing accordingly.
The share of unemployed people in occupations with a high risk of leaving the job and the number of job openings in occupations with a high risk of leaving the job include software developers and telemarketers. Among software developers, AI is responsible for 12% to 20% of the increase in unemployment recorded between 2022 and 2024 and 2025. Among salespeople, it’s described as between 10% and 26% of the increase.
In both cases, the researchers explained, the impact reflects not only a reduction in the number of vacancies, but also changes in required skills, making it harder to match unemployed people to available jobs.
The researchers said that the rise in unemployment rates in occupations exposed to AI is driven not only by job disappearances, but also by a widening mismatch between workers and available positions. In occupations where vacancies are not decreasing at the same pace as unemployment rates are rising, displaced workers are likely to find it increasingly difficult to integrate into existing jobs due to changing requirements from employers.
The research team noted that this trend is not solely driven by AI. “This also reflects structural factors such as a slowdown in high-tech sectors, a growing proportion of digital-age occupations at risk of automation for both employed and unemployed workers, and a partial retreat from the structural changes brought about by the coronavirus crisis. This study suggests that compared to these other factors, AI’s share of overall change is still modest,” the researchers wrote.
Debowy said jerusalem post The process by which robots put people out of work will be slower than generative AI due to the higher implementation costs of their use, but will ultimately have an impact. In Israel, robots have reduced the need for traditional local manufacturing workers, replacing a third of such workers in recent years. ”
Asked about the work of civil servants, Mr De Bowie said the sector is “slow to react because it depends on the decisions of political leaders rather than market forces. We don’t know where the civil service will go.”
Debowy, a Taub Center researcher who graduated from Columbia University with a bachelor’s degree in economics, is pursuing a doctoral degree at Ben-Gurion University of the Negev (BGU) in Beersheba.
There is a growing preference for more experienced workers. In reality, AI could significantly increase the productivity of experienced, highly skilled workers, shifting demand away from workers early in their careers. The findings are consistent with evidence from the U.S. that employment among younger workers (22 to 25 years old) in jobs at risk of automation fell by 13%, while experienced workers were largely unaffected, the researchers wrote.
The data shows that even in jobs where the number of job openings is relatively stable, the number of skilled unemployed people competing for those positions is increasing, making each position more competitive. This has increased the pressure on job seekers, who now need to present higher levels of experience and skills than before to secure a job.
One piece of good news is that while overall unemployment rates have remained stable, the nature of the labor market is changing. The researchers emphasize that, “Although the impact of AI on Israel’s overall unemployment rate remains modest, the changes it is causing cannot be ignored. Unemployment has not increased in the AI era, but the composition of the unemployed has changed. Between 2 and 6 percent of the changes in the occupational distribution of the unemployed are already explained by AI.”
People replaced by AI will often be able to improve their skills or even find new jobs if needed, he added. “Some people will lose their jobs, but others will get new jobs. Some will benefit and others will lose. The situation is fluid. There will be a new balance,” Debowy suggested.
As for the impact on teachers, he said, “We looked into it, but we didn’t find that AI was causing the employment of educators, because children would be taught by virtual teachers on screens. There is currently a severe shortage of teachers in Israel.”
