AI and the American Contracting State

• In Event: AI in Public Procurement, Budgeting, and Government Operations

Abstract

In the United States, many key governance decisions are made at the state and local levels, from housing to education to criminal justice. These are the sites where policy is translated into practice and where citizens most directly encounter the state. Artificial intelligence is transforming these implementation frontiers, yet state and local governments rarely build AI systems themselves. Instead, they procure AI technologies from private vendors. This article introduces the concept of the American AI contracting state to describe how public authority over AI is exercised through procurement. It shifts attention away from federal regulation and technical design, and toward the contractual relationships that structure how AI systems enter and operate within government. Drawing on more than 600 state AI contracts obtained by the Electronic Privacy Information Center and over 1,000 AI-related procurement reports from StateScoop, I map the emerging political economy of state-level AI markets. I show that public sector AI diffusion is contract driven, concentrated among a small ecosystem of vendors, and largely shielded from public visibility. The median contract value is approximately one million dollars, and the average duration is seven years, indicating long-term institutional commitments. These contracts often bundle software, data infrastructure, and ongoing services, which can create dependencies that are difficult to unwind. The most commonly contracted AI products are fraud detection and risk classification systems. These systems are frequently deployed in policy domains such as public benefits, unemployment insurance, and child welfare. In these settings, they expand the state’s monitoring and enforcement capacity through private intermediaries. This arrangement raises important questions about accountability, as decision-making authority becomes distributed across public agencies and private firms. A substantial share of these contracts is not reported, even by specialized government IT media, which further limits public oversight. Procurement therefore functions as a hidden but consequential governance mechanism that structures service delivery, shapes administrative discretion, and redistributes power between the state and private actors. Studying AI governance in public policy requires analyzing contracting not as a site of political contestation over values such as equity, transparency, and democratic control.

Author

• Jae Yeon Kim, University of North Carolina at Chapel Hill
  Presenting Author