Which Jobs Are Most Threatened by AI and Who Can Adapt

New research from GovAI and the Brookings Institution offers a novel perspective on how artificial intelligence (AI) could reshape the workforce, identifying jobs most threatened and those best positioned to adapt. While many workers with high AI exposure are also those most likely to find new opportunities, the study highlights a vulnerable segment: a significant number of largely clerical and administrative workers, predominantly women, who face both high AI risk and low adaptability for new roles. This analysis arrives amidst conflicting predictions about AI's future impact, aiming to provide a clearer, albeit still fallible, understanding for workers and policymakers.

Unpacking AI's Conflicting Job Forecasts

The discussion around AI's impact on the labor market is rife with contradictions, leaving many dizzy with uncertainty. Some analyses suggest AI is already displacing young workers in fields like software development and customer service, while others claim young workers in these very fields are thriving. Forecasts range from AI posing no significant job threat in the next decade to prominent CEOs predicting millions of job losses imminently. This divergence underscores a significant gap between what is known about AI's effect on work and the widespread desire for definitive answers.

Despite the conflicting narratives, two points generally coalesce among economists: there is currently no measurable evidence of AI causing widespread job displacement across the American workforce. Historically, automation primarily affected factory and trade jobs, but today, white-collar occupations are at the forefront of AI-driven transformation.

A New Lens on AI Exposure and Adaptability

Researchers Sam Manning of GovAI and Tomás Aguirre developed a unique approach to estimate AI's labor market impact. They began by quantifying AI "exposure" for over 350 occupations, assessing how many job tasks could be made more efficient by AI, such as a teacher grading homework. This initial assessment revealed high overlap between AI capabilities and skills used in computer programming, marketing, financial analysis, and customer service, theoretically making these workers more susceptible to replacement.

However, Manning and Aguirre extended their analysis by also measuring workers' potential to adapt to new jobs if AI rendered their current roles obsolete. Their model factored in education level, varied work experience, wealth, age (under 55), and residence in cities with abundant job opportunities as key indicators of adaptability. This dual approach provides a more nuanced view beyond simple exposure.

The Most Vulnerable: Clerical and Administrative Roles

The research unveiled a significant divergence between high AI exposure and adaptability. While occupations like web designers and secretaries both showed high exposure, their adaptability scores varied dramatically. Secretaries, along with 6.1 million other clerical and administrative workers, were identified as having both high AI exposure and among the lowest estimated adaptability. The study starkly highlights that approximately 86 percent of these most vulnerable workers are women, suggesting an unequal distribution of AI's negative societal effects.

Allison Elias, a University of Virginia business school professor not involved in the study, points to historical patterns. In past technological shifts, women in clerical roles often hoped new tools would elevate their work, only to find themselves doing more for the same or less pay, with consistently low job satisfaction. Elias emphasizes that these workers are "really vulnerable because they won’t have a lot of decisions over how AI is used, and their exit opportunities are going to be pretty low."

Lessons from History: Predictions Often Miss the Mark

Economists caution against overly relying on current AI capabilities or early adoption trends to predict its long-term labor market effects. History is replete with examples of technological revolutions, like electricity and smartphones, that eliminated certain jobs while simultaneously creating entirely new industries and occupations that were unforeseen.

For instance, a prominent study over a decade ago wrongly predicted nearly half of all jobs would be destroyed by computer automation. Similarly, forecasts that ATMs would eliminate bank tellers, earlier AI forms would decimate radiologists, and player pianos would end pianists' careers proved unfounded. Few could have imagined smartphones ushering in roles like social media marketing and influencing. Martha Gimbel of Yale University's Budget Lab aptly notes, "We do not have a good track record of predicting how technological change will play out in the labor market."

Hope and Pessimism from Extinct Occupations

The demise of telephone switchboard operators, once a common job for American women, offers a mixed historical lesson. As telephone technology advanced in the early 20th century, these jobs were largely eliminated. Research by James Feigenbaum and Daniel Gross found that displaced switchboard operators were significantly more likely to remain jobless or accept lower-paying work. However, within years, new opportunities boomed for young women in secretarial and restaurant work. Feigenbaum, an economic historian, sees this as "somewhat hopeful."

Feigenbaum argues that AI will likely follow a similar trajectory to past transformative technologies like electricity or the internet – massive changes, but not the elimination of all jobs. The consensus among experts remains one of "humility," stressing that while AI will undoubtedly transform work, its ultimate shape and societal impact are still largely unwritten.

FAQ

Q: Is AI already causing widespread job loss in the United States?

A: Economists currently find no measurable evidence that AI is causing widespread job displacement across the American workforce as a whole. However, AI is beginning to cause shake-ups, particularly in white-collar occupations.

Q: Which types of workers are identified as most vulnerable to AI by new research?

A: Recent research from GovAI and the Brookings Institution suggests that clerical and administrative workers, such as secretaries, are among the most vulnerable. This group, disproportionately women (86%), faces high AI exposure coupled with low estimated adaptability to new job roles.

Q: How accurate have past predictions about technology's impact on jobs been?

A: Historically, economists and researchers have had a poor track record in accurately predicting the effects of new technologies on work. Past forecasts, such as ATMs eliminating bank tellers or earlier AI forms decimating radiologists, proved largely incorrect, as new technologies often create unforeseen jobs while displacing others.