Regulating Autonomous AI Agents: Moral Agency, Liability, and Governance in Self-Evolving Systems

Abstract

Autonomous Artificial Intelligence (AI) agents have transitioned from decision-support tools to self-evolving digital entities capable of independent reasoning, learning, and goal-directed execution. This transformation introduces unprecedented regulatory dilemmas regarding moral agency, accountability, liability, and governance. Traditional legal frameworks grounded in human intent and corporate responsibility struggle to accommodate systems that generate emergent behaviors beyond explicit programming. This research investigates the intersection between AI autonomy, ethical agency, and regulatory oversight, proposing a hybrid governance model that integrates technological accountability layers, explainability mandates, dynamic compliance protocols, and adaptive liability distribution frameworks. Using a comparative legal analysis across the European Union, United States, China, and India, combined with case review of real-world AI failures, the study synthesizes actionable future-ready policies for regulating self-evolving AI agents. Findings reveal that meaningful oversight requires redefining AI moral agency, restructuring accountability chains, and embedding algorithmic compliance directly into AI system architectures