Agents in a Tangled Bank: An Ecosystem Approach to AI Regulation

As autonomous artificial intelligence agents ("Agents") become increasingly prevalent in society, legal frameworks must evolve to govern their behavior effectively. This article argues that autonomous AI agents must be understood as operating within complex ecosystems of other agents, humans, institutions, and actual biological ecosystems-similar to how biological organisms exist within broader ecosystems. Drawing on established research in multi-agent systems and environmental law, we propose that effective governance of AI agents requires moving beyond just regulation on individual agents to include system-level approaches. We examine how existing legal frameworks handle heterogeneous autonomous entities (humans, corporations, and animals) and explore how these frameworks might extend to AI agent ecosystems.

Finally, we propose a set of concrete principles for ecosystem-based AI governance. This approach offers more effective ways to manage agent interactions, prevent systemic risks, and promote beneficial outcomes in increasingly complex multi-agent environments.