AI Impact on Job Market

The question of whether artificial intelligence will take jobs is being widely considered. New research by Bryan Seegmiller and Dimitris Papanikolaou at Kellogg hints at how the labor market will be affected by AI. The researchers and their colleagues studied how workers’ jobs are exposed to AI and the impact on employment across industries. Their findings suggest the effect of AI on the labor market over the last decade is not straightforward.

The study found that the more a job is exposed to AI, the more likely the demand for that job will decrease. However, workers in jobs with higher AI exposure could adjust by focusing on tasks with less AI exposure and improving in those areas. Also, companies that use AI extensively tend to see increased productivity and workforce expansion. The change in demand for highly paid positions was relatively flat, even with high AI exposure, because these opposing effects balanced out.

Seegmiller suggests that workers may need to change their responsibilities to tasks that work with AI to survive the AI boom. Focusing more on communication, big-picture thinking, and collaboration might be helpful. Flexibility will be important “to mitigate its negative effects,” Seegmiller notes.

Seegmiller and his colleagues had previously studied how technological advances affected jobs. They found that mid-level wage occupations were negatively affected as robots, software, and IT disrupted industries and decreased demand for those workers. Some people had difficulty adapting to new workplace requirements. Seegmiller says that “this type of reallocation can be really painful.”

Seegmiller and Papanikolaou, along with Menaka Hampole and Lawrence D.W. Schmidt, researched whether similar upheavals could happen with the rise of AI. Seegmiller says AI will significantly shape the labor market over the next few decades and that if a company used AI heavily, the firm tended to increase its overall productivity and expand its workforce.

LinkedIn Data Analysis

The team analyzed approximately 58 million LinkedIn profiles from Revelio Labs, focusing on U.S. jobs held between 2014 and 2023. They determined how companies used AI for specific functions based on resume information. For example, one J.P. Morgan employee used AI software to forecast risk and fraud. The researchers compared the tasks AI performed with tasks in O*NET, which provides information on job duties and skill requirements. If an AI function resembled a task done by people, that task was considered “exposed” to AI, meaning AI could replace a human worker.

The team discovered that AI exposure was higher in higher-paying, white-collar jobs, peaking at the ninetieth percentile of income. Occupations with high exposure included financial specialists, life-science technicians, chemical engineers, and credit analysts. Jobs involving manual labor, like bartenders, janitors, and cooks, had the least exposure.

The researchers’ model indicated a more complex relationship between AI and employment. Jobs with a wide range of tasks that have different levels of exposure to AI, or high “variance,” reduced the likelihood of displacement because workers could adjust their responsibilities. For example, workers could spend more time strategizing or forming business relationships if AI replaced routine tasks.

Seegmiller used AI to code an economic model, which saved him several hours. As a professor, his job's high variance gave him other work options, like writing papers. He said that the ability to “focus on things that I’m now more productive in, because I don’t have to spend time on other stuff, is actually good for me.”

AI Adoption and Employment

Furthermore, a firm's AI adoption level also affected employment. The team measured AI adoption using LinkedIn data by analyzing how often employees mentioned AI. They found that AI-intensive firms saw increased productivity, enabling them to expand their workforce.

Accounting for all these factors, the net effect of AI on employment was close to zero, especially for highly paid jobs. Although AI replaced some tasks, these jobs often had high variance, allowing workers to shift to other tasks. Also, these workers were more likely to be at firms that used AI enough to increase productivity and employment growth. “Employment share” saw a slight increase for jobs at the very top of the income scale.

Seegmiller says that the effect of AI “then, ultimately doesn’t just depend on whether some of your tasks get automated,” but on “the sum of these forces.”

The net effect on employment share was still negative for some highly paid jobs, including those in business, finance, and engineering. Business and financial occupations saw a 1.9 percent decrease in employment share over five years, and architecture and engineering saw a 2.6 percent drop. Some lower-paid, manual jobs also saw declines in employment share. Food preparation and serving occupations, for example, saw less AI-driven growth, resulting in a 2 percent decrease in employment share.

AI was a significant factor in changes to the labor market. AI-related factors explained approximately 14 percent of the changes in employment growth across all occupations during the study period.

Future of AI and Work

Advancements in large-language models suggest AI’s role in society will continue to grow. Legal occupations, which involve sifting through large amounts of text, might see more tasks automated, Seegmiller says, and software engineers might offload writing code to AI. The outcome for these jobs will depend on worker flexibility and company decisions about AI use. Engineers could adapt by shifting to more high-level strategy instead.

As workers prepare for AI-driven changes, they may want to reevaluate their jobs and consider working “in conjunction with AI rather than in competition with AI,” according to Seegmiller.