NEW YORK — The widespread anxiety that artificial intelligence (AI) will cause mass job displacement is likely exaggerated, according to a new report from the Yale Budget Lab (also known as the Economic Growth Center at Yale). The study, released on October 1st (local time), compared the impact of generative AI adoption with the earlier spread of the internet, concluding that the immediate disruption to the U.S. labor market is far less severe than feared.

Current AI Impact vs. the Internet Boom 

The Yale researchers analyzed U.S. employment changes over the 33 months since the launch of ChatGPT in November 2022, benchmarking them against the period of internet and computer diffusion between 1996 and 2002.

While the rate of change in the occupational composition has accelerated since the introduction of generative AI, the increase is modest—only about 1 percentage point faster than during the early internet era. Crucially, the researchers noted that this acceleration trend in occupational change was already underway prior to the generative AI boom, making it difficult to attribute the entire shift solely to AI.

During the initial phase of internet commercialization (1996–2002), the change in the composition of U.S. jobs amounted to about 7%. The report suggests that the current pace of change due to AI is not a significant leap, challenging the notion of a sudden, catastrophic shock to employment.

No Evidence of Job Loss in AI-Exposed Roles 

Contrary to popular belief, the study found no significant evidence of a decline in employment for job categories highly exposed to AI. Occupations deemed "highly exposed"—where AI could reduce work time by more than half—have maintained a relatively stable proportion of the workforce (18% in 2022 to 18% in August 2025). The proportions for moderately-exposed and low-exposed roles also saw little to no change. Furthermore, the employment trajectories of recent college graduates (ages 20–24) and older graduates (ages 25–34) showed no marked divergence attributable specifically to AI, suggesting that current anxieties are largely speculative rather than data-driven.

Long-Term View and Quality of Employment 

The report emphasizes that major technological transformations, like the full industrial impact of personal computers, typically take over a decade to fully materialize. AI's long-term effects on the labor market will likely unfold over a similar extended period.

Furthermore, some experts argue that the critical issue is not the sheer quantity of jobs, but their quality and distribution. Lee Sang-hun, Director of Employment Policy at the International Labour Organization (ILO), stated that the real challenge posed by technological change has historically been about job quality and distribution, which are ultimately matters of policy, politics, and investment in people.

The Yale study offers a dose of realism, suggesting that while AI is undoubtedly transformative, the immediate employment shock is overstated, and current data reflects stability over economy-wide disruption.