Artificial Intelligence, Regulatory Capacity, and the Adaptive State

ABSTRACT Artificial intelligence is changing not only what regulators govern but the informational environment through which regulation is conducted. Regulators increasingly use advanced analytics to process evidence and target supervisory attention, while regulated organizations deploy AI across their operations. This paper develops the concept of the adaptive regulatory state to analyze how changing informational capabilities generate behavioral and institutional adaptation. Comparative cases from financial supervision, competition, tax administration, and electronic communications identify three possible trajectories: constructive learning, governed co‐evolution, and adversarial escalation. Effective oversight does not require technological parity with regulated firms, but sufficient informational, organizational, and verification capabilities to respond as the regulatory environment changes. Four institutional principles follow: functional rather than technological capacity; effective inspectability and proportionate assurance; learning with bounded revisability; and coordinated but accountable capacity. The adaptive regulatory state is defined by whether regulatory institutions remain capable, accountable, and legitimate as their informational environment evolves.