Governing generative AI under uncertainty: institutional anchoring and regime formation in urban school districts

This paper examines how large urban school districts in the United States govern generative artificial intelligence (GenAI) in the absence of comprehensive federal or state regulatory mandates. Drawing on a comparative analysis of twenty-nine governance artifacts produced between 2023 and early 2026 across four major districts (Chicago Public Schools, New York City Public Schools, the School District of Philadelphia, and Miami-Dade County Public Schools), the study identifies four distinct AI governance regimes anchored in different institutional logics. Chicago constructs an infrastructure-embedded regime that integrates AI governance into administrative modernization and compliance systems. New York anchors authority in statutory transparency mandates and algorithmic reporting requirements. Philadelphia governs through professional mediation and bounded vendor authorization. Miami represents a formation-stage regime in which governance structures remain under deliberative construction. The study advances a policy-centered argument: under conditions of regulatory uncertainty, districts do not merely adopt or restrict AI; they construct provisional governance regimes that stabilize the institutional meaning of AI in public education. These regimes reflect distinct theories of control, accountability, and institutional responsibility. Their differences carry equity implications, distributing risk and authority unevenly across educators, students, and vendors. The findings contribute to educational policy scholarship by theorizing district-level AI governance as institutional anchoring under uncertainty and demonstrating how early local decisions may entrench or mitigate disparities before higher-level regulatory frameworks solidify.