Technological advancements and innovation have historically sparked anxiety among workers. For example, during the early 1800s, the Luddites famously destroyed textile machinery in protest against automation that threatened their jobs. Similarly, economist John Maynard Keynes suggested in 1930 that technological progress could lead to temporary mass unemployment. Today, similar concerns persist about robots and artificial intelligence replacing human workers. While it’s easy to highlight the displacement effects technology can have on labor, it’s equally important to recognize that innovation can also increase demand for labor in new industries, ultimately resulting in a net positive economic impact.
The historical relationship between innovation and the labor market is complex. In some instances, technological progress has created new opportunities and boosted labor demand, while in other cases, it has displaced workers. This raises the crucial question: How will generative AI impact labor markets, and what will its broader economic implications be?
A review of the economics of innovation helps shed light on how technological advancements interact with the broader economy. Research spanning several decades shows that, in the United States, innovation was closely tied to stronger labor demand in the four decades following World War II. However, in the subsequent four decades, innovation has led to comparatively weaker labor demand. This does not mean that labor demand has decreased, but rather that it has been less robust than in the earlier period. Notably, the employment-to-population ratio in the U.S. peaked in 2000, well into this slower growth phase for labor demand .
Understanding this shift highlights how innovation can impact labor markets in different periods, with complex outcomes that vary by the nature of the technological changes. While innovation has historically been a driver of labor demand, the effects on the economy may not always be linear or uniformly positive, especially as new technologies alter existing industry dynamics and job structures.
The stronger economic growth observed in the decades following World War II is often attributed to the significant impact of the “Great Inventions of the Second Industrial Revolution.” These include transformative innovations such as electricity (e.g., the light bulb and electric motor), the internal combustion engine, sanitation (e.g., running water and indoor plumbing), chemicals (e.g., natural gas, plastics, and pharmaceuticals), and telecommunications (e.g., the telephone and radio). Researchers suggest that the profound effects of these inventions on both the economy and living standards were so exceptional that repeating such large-scale innovations may be unlikely in the futurenovations of the post-World War II period primarily consisted of product innovations—new or significantly improved goods and services—rather than process innovations, which focus on enhancing the methods of production. For example, the invention of the automobile is a product innovation, whereas industrial robots used in car manufacturing represent process innovations. Product innovations are generally more likely to improve productivity and stimulate economic growth compared to process innovations .
From rket perspective, innovations that augment human labor or handle tasks humans cannot perform are typically seen as contributing to stronger labor demand and economic growth. These innovations align more closely with product innovations. In contrast, labor-automating technologies, like AI, that focus on replicating existing human tasks are more akin to process innovations and may reduce overall labor demand .
However, distibetween product and process innovations or labor-augmenting and labor-automating technologies is often not straightforward. For instance, the automobile is both a product innovation (in its market appeal and function) and a process innovation (in its impact on logistics). Similarly, technologies like word processors may be labor-automating for a secretary but labor-augmenting for a lawyer .
Moreover, the scale of plays a significant role in determining its economic impact. Research suggests that “radical innovations” tend to have a more substantial positive effect on productivity growth compared to “incremental innovations” . For example, the invention of ref had a transformative impact on agriculture and food processing by reducing spoilage, whereas a “so-so” innovation, such as a self-service kiosk that shifts work from a cashier to a customer without improving service quality, may lead to small productivity gains but not result in substantial economic benefits .
From a theoretical perspective, the impact of innovation on the labor market can be understood through three key effects: the displacement effect, the reinstatement effect, and the productivity effect.
- Displacement Effect: This occurs when innovation automates existing tasks, leading to a reduction in labor demand. As certain tasks become automated, fewer workers are needed to perform them.
- Reinstatement Effect: On the other hand, innovation can also create new tasks, boosting labor demand. For instance, the rise of automation technologies can increase the need for data scientists and computer specialists to develop, manage, and maintain these technologies.
- Productivity Effect: Innovation can increase labor demand in industries that are unaffected by automation. This happens when productivity improvements in one sector lower costs and allow businesses to invest those savings into other areas, driving growth and, in turn, demand for labor. For example, the introduction of refrigeration lowered food spoilage, reducing costs across industries, which led to labor demand growth in sectors benefiting from these savings.
Despite the preference for labor-augmenting product innovations over labor-automating process innovations, experts argue that the magnitude of the productivity effect is the most crucial factor in determining whether innovation will ultimately increase or decrease overall labor demand. Innovations that lead to large cost savings, especially when wages are high and labor is scarce, are more likely to drive stronger productivity growth, thereby boosting labor demand across the economy. This highlights how the productivity effect is most potent in sectors where wages are elevated, and labor shortages are more pronounced.
Generative AI, a technology that creates new content such as text, images, code, and voice using foundation models, is still in its early stages. However, its potential to impact labor markets can be assessed based on current trends and predictions. One important context for understanding its potential effect is the historical productivity outcomes of innovation, particularly the weaker effects seen over the past 40 years.
In the past, innovation had a more limited productivity effect, particularly because automation mainly targeted low- and middle-wage workers. For example, tools like word processors and spreadsheets automated clerical tasks, while machinery and industrial robots took over factory work. These workers were often in low- or middle-wage jobs, and the savings from automation were relatively modest, leading to smaller productivity effects.
Generative AI, however, may have a different impact. Research suggests that high-wage workers are more vulnerable to being replaced by this technology than low-wage workers. Occupations such as postsecondary educators, mathematicians, and survey researchers, many of which are high-wage, are at risk of being automated by generative AI. The industries most affected include legal, financial, and professional services. Since these sectors tend to involve high wages, the cost savings from automating tasks could be substantial. If these savings translate to larger productivity effects, it is likely that generative AI will lead to stronger aggregate labor demand, rather than weaker demand, despite the displacement of high-wage workers. This shift could create opportunities in other areas of the economy, potentially driving growth in sectors not directly impacted by AI automation.
In conclusion, while generative AI may replace certain tasks currently performed by high-wage workers, the cost savings and productivity gains could ultimately drive stronger labor demand, especially in industries that are less vulnerable to automation.
Generative AI, while often categorized as a process innovation, has the potential to create mixed outcomes for labor demand. Process innovations, typically those that improve how tasks are performed, have a lower likelihood of boosting labor demand compared to product innovations, which introduce entirely new goods or services. However, the impact of a process innovation depends on whether it’s a “brilliant” or “so-so” technology. “Brilliant” technologies, such as refrigeration or electricity, lead to large productivity gains and subsequently higher labor demand. Generative AI, due to its wide-ranging applications across industries and its ability to automate complex tasks, is considered closer to a “brilliant” technology, which bodes well for strong productivity effects.
Generative AI’s potential to increase labor demand is influenced by two key factors: demographics and institutional frameworks. First, the scarcity of labor, particularly in aging populations in developed countries, may encourage firms to adopt generative AI to counteract labor shortages. This adoption could lead to significant productivity gains, especially if the cost savings are substantial.
The second factor is institutional. Labor protections and union presence can significantly mitigate the displacement effect of generative AI. In countries with strong worker protections, the labor market may experience a more gradual shift, with workers being reassigned to new roles within the same organizations rather than displaced. For example, in Germany, the introduction of industrial robots did not significantly reduce local labor demand, as workers were able to transition within their companies. Stronger labor protections may therefore dampen the negative effects of generative AI, leading to a smoother transition and potentially increased labor demand in the long run.
Overall, while generative AI may displace certain tasks, its broader productivity effects, especially when combined with favorable demographic and institutional conditions, suggest that the technology could ultimately support stronger labor demand.