A new study from Federal Reserve economists has quantified the impact of generative AI on the tech labor market, finding that employment growth among US programmers was cut in half following the launch of ChatGPT.
"This is the first Federal Reserve-level study to directly link AI adoption to a measurable, occupation-specific decline in developer hiring," Fed economists Leland D. Crane and Paul E. Soto wrote, pointing to AI as the cause of an "occupation-specific shock."
Before November 2022, programming jobs grew around 5% annually. The study found that growth has since flatlined in tech-heavy sectors, creating an employment gap of roughly 500,000 jobs over three years compared to pre-AI trends. The researchers noted the hiring slowdown did not appear in the data until mid-2024, suggesting companies took time to evaluate the technology before adjusting headcount.
The findings add institutional weight to concerns about AI's effect on white-collar jobs, particularly for junior talent. A separate Harvard study found that junior developer employment drops 9-10% within six quarters when companies adopt generative AI, raising long-term questions about the career pipeline and income disparity.
Counter-Narrative Emerges
Despite the Fed's findings, some industry leaders are actively increasing entry-level hiring. Salesforce CEO Marc Benioff announced the company is hiring 1,000 new graduates and interns specifically to build its AI systems, directly countering the narrative that AI eliminates such roles.
IBM is making a similar bet, tripling its hiring for entry-level positions, including software development. "The companies three to five years from now that are going to be the most successful are those companies that doubled down on entry-level hiring in this environment," Nickle LaMoreaux, IBM’s chief human resources officer, said in an interview.
Augment or Replace
The debate centers on whether AI will primarily augment or replace workers. The Fed study shows programmers are the most AI-exposed group, and data from Anthropic's Claude.ai shows coding-related tasks make up nearly half of its enterprise API traffic.
While firms like Block have cut staff citing AI efficiency, the broader macroeconomic data has yet to show mass unemployment. Nvidia CEO Jensen Huang argues that AI agents will work alongside employees, boosting productivity without reducing headcount. The Fed study itself notes that programmer wages have not declined, and cheaper AI-assisted development could eventually open new markets and increase overall demand for labor.
This article is for informational purposes only and does not constitute investment advice.
