Fernando Brito

University of Florida · Food and Resource Economics · 2026–27 job market

Fields

Labor, Agricultural, Innovation, Econometrics, Environmental & Energy, Urban & Regional

Job market paper

Artificial Intelligence (AI) and Agricultural Employment

We quantify how artificial intelligence (AI) innovations affect U.S. agricultural employment. Using 2003–2023 U.S. Patent and Trademark Office records, we construct a measure of agriculture's exposure to AI by classifying AI patents and mapping them to subsectors through Cooperative Patent Classification (CPC) and North American Industry Classification System (NAICS) concordances. A shift–share design with an instrumental variable approach interacts national AI innovation trends with county-level lagged employment shares. Estimates indicate that AI exposure raises employment in crop production (NAICS 111) by 10.6% and in animal production (NAICS 112) by 6.8%, relative to mean employment levels in each subsector. Decomposing AI based on its underlying function reveals heterogeneity in employment effects. Execution-oriented AI (hardware and automation) reduces overall agricultural employment (NAICS 11) by 6.9%, while cognition-oriented AI (decision-support) and perception-oriented AI (sensing) increase employment (4.3% and 6%, respectively). These domain-specific patterns are consistent across crop and animal subsectors, with stronger employment effects in crop agriculture (-12.8% for execution-, 9.1% for cognition-, and 10% for perception-oriented AI) than in animal agriculture (-4.5% for execution-, and 7.8% for perception-oriented AI, respectively). Results show that AI does not uniformly displace farm labor; instead, its effects depend on technological complementarities and task reallocation. By uncovering domain-specific dynamics, this study demonstrates how targeted AI innovations can augment agricultural employment, informing policies that support skill upgrading and diffusion of complementary technologies.