Andrew W. Torrance (University of Kansas School of Law; MIT Sloan School of Management) & Bill Tomlinson (University of California, Irvine; Victoria University of Wellington – Te Herenga Waka) have posted Secret Agents: Hidden Governance of Autonomous Artificial Intelligence (European Journal of Law Reform (accepted)) on SSRN. Here is the abstract:
This Article examines how the historical arc of autonomous software agents, beginning with ancient Greek automatons and culminating in modern large language models, can guide the governance of artificial intelligence in contemporary society. It explores decentralization, emergent behavior, potential for surprise, and assumptions of beneficence as cornerstones of effective regulatory strategies. The discussion highlights that modern principles of AI governance may not fully capture these underappreciated dimensions, particularly as large language model agents evolve beyond simple task execution to serve as orchestrators of problem-solving in multiple fields. The text proposes a recalibration of legal structures, urging policymakers and scholars to look to the past for insights on how best to manage and channel the power of increasingly independent AI agents.
Highly recommended.