The Federal Reserve has confirmed a significant decline in U.S. programming job growth, with a study revealing that job openings for programmers plummeted by approximately 50% following the launch of ChatGPT in November 2022. Researchers Crane and Soto highlight that an estimated 500,000 developer positions that would have otherwise existed went unfilled, as the employment gap emerged about 18 months after the AI’s introduction. Notably, the study identifies programmers as the most affected occupational group due to AI adoption, particularly junior developers whose roles saw sharper declines compared to senior positions, raising concerns about the future of career pipelines and mobility within the tech sector.

Paul E. Soto: Paul E. Soto is a principal economist in the Industrial Output Section of the Federal Reserve Board, with research interests in financial economics, financial intermediation, and artificial intelligence. He collaborated with Leland D. Crane on the March 2026 paper ‘AI and Coder Employment: Compiling the Evidence,’ which attributes slowed US programmer employment growth to large language models using industry counterfactuals. Soto’s work explores generative AI’s broader implications for labor and productivity.
Federal Reserve: The Federal Reserve is the central banking system of the United States, responsible for conducting monetary policy, regulating banks, and conducting economic research through divisions like Research and Statistics. In March 2026, its economists published a preliminary study titled ‘AI and Coder Employment: Compiling the Evidence,’ which provides the first institutional analysis linking large language model adoption to slowed growth in US programmer employment. The research uses counterfactual methods to isolate AI’s effects from factors like interest rate hikes and the post-pandemic slowdown.
Leland D. Crane: Leland D. Crane is a principal economist in the Industrial Output Section of the Federal Reserve Board’s Research and Statistics division, specializing in nontraditional data sources, artificial intelligence, labor markets, and search-and-matching models. He co-authored the March 2026 Federal Reserve study ‘AI and Coder Employment: Compiling the Evidence’ with Paul E. Soto, documenting an AI-driven decline in programmer job growth starting mid-2024. His recent contributions include prior work on measuring AI uptake in workplaces.

AI Occupational Shock: The study identifies programmers as the most AI-exposed occupational group, with effects emerging about 18 months after ChatGPT’s launch as firms integrated LLM capabilities.
Counterfactual Analysis: Economists constructed industry-level counterfactuals to separate AI impacts from tech sector challenges like rate hikes and the crypto downturn.
Junior Developer Impact: Generative AI adoption correlates with sharper declines in junior developer roles compared to senior positions, raising concerns for career pipelines and mobility.