Translated from Korean

Youth employment in AI-exposed jobs must be viewed alongside actual adoption records

If youth hiring has cooled first, it’s worth suspecting the impact of AI. Stanford researchers analyzed U.S. ADP payroll data through June 2026 and found that employment for 22-25-year-olds was weaker in occupations more exposed to AI compared to those less exposed. This difference was mainly reflected in new hires. However, there were already some trends in differences before the widespread adoption of generative AI, and the researchers did not interpret these results as a causal effect of AI adoption.

In Denmark, early-career employment also declined in some AI-exposed occupations. However, when researchers compared firms based on whether they had actually adopted chatbots, there was no evidence that the proportion of early-career staff decreased more in workplaces that implemented them. The Danish results cannot explain the causes of the hiring slowdown in the U.S. Instead, this case demonstrates that even when employment changes appear similar on the surface, interpretations can shift once actual adoption records are applied.

In the U.S., it would be useful to compare new hiring around the time companies adopted chatbots for the same roles against companies that have yet to do so. It is necessary to verify the point at which AI became capable of performing specific tasks and the point at which companies actually reduced their hiring separately.

Loading comments…