Does AI Really Threaten Office Jobs?

AI and the Fear of Massive Disappearance of Office Jobs

Recently, the tech sector has been shaken by a series of massive layoffs affecting renowned companies such as Coinbase, Meta, and Cisco. These events have fueled a growing fear that artificial intelligence (AI) is on the verge of destroying office jobs. However, before making drastic decisions regarding your career, it is crucial to examine current economic research on the actual impact of AI on these jobs.

The answer to this concern is rather reassuring: No, AI does not yet seem to have a significant impact on the American labor market. Despite alarming predictions of a job apocalypse, concrete evidence is lacking to support the idea that AI has already caused massive changes.

Data from the U.S. Bureau of Labor Statistics (BLS) reveals that the unemployment rate in jobs potentially most vulnerable to AI is actually lower than in other less exposed professions. Moreover, there is no massive movement of workers leaving jobs threatened by AI for positions considered safer, such as those primarily involving manual labor.

While current statistics do not guarantee that the labor market will not undergo upheavals in the future, they challenge the inevitability of catastrophic scenarios and the speed at which they might occur. In the world of AI, many predict that technology will soon obliterate jobs, but current data paints a picture of a relatively stable labor market.

Current Data and Its Interpretation

Erika McEntarfer, a labor economist, emphasizes that available evidence indicates that the impact of AI on the labor market is still weak. A former director of the BLS, she was dismissed by President Trump after a jobs report displeased the administration. Since her dismissal, BLS reports on slow job growth have continued.

McEntarfer, who is now a researcher at the Stanford Institute for Economic Policy Research, asserts that history shows innovations take time to transform industries and professions. According to her, AI will not disrupt labor markets until it has first transformed businesses.

Data from the U.S. Census shows that one in five businesses uses AI in some commercial function. McEntarfer insists that this data helps to put the fear of an imminent massive disruption into perspective. While AI may be disruptive in the future, current data indicates that this disruption is not yet here, allowing time for planning.

The American Labor Market and Recent Graduates

The labor market in the United States remains challenging for many young people, particularly for those who have just graduated. The unemployment rate for recent graduates hovers around 5.6%, a high figure compared to the overall workforce. This rate recalls levels observed during the pandemic and after the 2008 recession.

Recent graduates, especially those seeking to enter the tech sector, face difficulties, but it is uncertain to what extent AI is responsible for this situation. Professions affected by AI represent only a small portion of the overall labor market, and other macroeconomic factors may also play a role.

There are signs that AI contributes to the challenges faced by young people aged 22 to 25 seeking jobs in software development and other professions heavily impacted by AI. However, these professions constitute only a fraction of the overall labor market. Therefore, it is difficult to determine the extent to which AI is responsible for these employment difficulties.

The Need for More Comprehensive Data

To better understand the impact of AI on the labor market, more comprehensive data is needed. Current statistics, derived from the federal government's monthly survey of 60,000 households, provide an overview but do not sufficiently explain how AI affects the diverse labor market in the United States.

Researchers, such as David Deming from Harvard University, point out that we lack data to answer crucial questions: How is AI being used in the workplace? Is it replacing workers or making them more productive? Which professions are most affected? Deming and his colleagues have surveyed several thousand people every three months since 2024, asking them about the use of generative AI.

Deming's results show that AI is used by just over 40% of workers, although adoption varies by sector. This allows for estimating productivity gains, which exist but are not revolutionary. This data helps document the speed at which AI has been adopted in the workplace compared to earlier technologies like PCs and the Internet.

Early Adopter AI Adoption

Studying early adopters of AI provides valuable insight into the future of the labor market. According to Deming, this helps to understand how AI might be used in the future and who might be most affected. However, it does not predict the fate of different jobs.

Gaining a picture of these early adopters and how they use AI provides a “crystal ball for the future of the labor market,” says Deming. “It gives you important clues about how this will be used tomorrow, who will be affected, who will be disadvantaged, and how we need to prepare for it. It's a diagnostic of what's coming.”

But what it does not tell you is the fate of various jobs.

Young Workers, the Most Vulnerable

Analyzing the impact of AI on jobs often begins by assessing the exposure of professions to technology. This approach is based on the idea that each job is a collection of tasks, and researchers evaluate which tasks can be performed by AI.

Studies classify jobs according to their vulnerability to AI, but these results do not necessarily predict job losses. This depends on many factors, such as AI adoption and economic calculations regarding the value of workers.

In a study conducted by the Stanford Digital Economy Lab, researchers examined 950 jobs and found a decline in the number of young workers in the most AI-exposed professions, such as software development, since the end of 2022. While other factors may explain these declines, researchers found compelling evidence of a significant effect of AI after 2024.

Researchers at the Stanford Digital Economy Lab used a vast dataset from ADP, the world's largest payroll provider, to examine job growth in different categories. Their exclusive access to this data allows for a better understanding of the impacts by demographics.

By analyzing the data, they discovered that young workers are particularly affected by these changes, while older workers appear less impacted. This trend could indicate that AI influences labor market dynamics in a more complex manner than initially expected.

They noted the decline in the number of workers aged 22 to 25 in the most exposed professions, such as software development and customer service, starting at the end of 2022, when ChatGPT was first released. Other researchers have reported evidence that the decline in these jobs began well before ChatGPT and have questioned how quickly the labor market might react to the introduction of AI technology.

However, while Stanford researchers acknowledge that other factors besides AI have likely contributed to the early declines, they assert that after controlling for these factors, they saw compelling evidence of a significant effect of AI after 2024, growing in 2025 to reach a 16% decline in entry-level jobs in AI-exposed professions. In contrast, the number of workers increased for older workers in the same professions, as did the number of jobs in less exposed professions.

By delving deeper into the data, researchers found another important clue.

Conclusion

Current data does not show massive disruptions caused by AI in the labor market, but it underscores the importance of remaining vigilant and continuing to collect data to better understand future developments.