Lessons from complexity theory for AI governance

TL;DR

The paper argues that AI systems exhibit properties of complex adaptive systems, including nonlinear growth, scaling, emergence, feedback loops, interdependence, and tail risks. It leverages complexity theory to analyze governance challenges under deep uncertainty and argues that traditional regulatory approaches may be ill-suited for rapidly evolving, interconnected AI ecosystems. It proposes three complexity-compatible governance principles—early and scalable intervention, adaptive governance, and complexity-aware risk thresholds—drawing on insights from climate policy and public health to guide proactive, information-driven regulation. The work highlights the practical importance of interdisciplinary collaboration and the need for regulatory frameworks that can adapt to rapid change and systemic interdependencies as AI technologies become more pervasive.

Abstract

The study of complex adaptive systems, pioneered in physics, biology, and the social sciences, offers important lessons for AI governance. Contemporary AI systems and the environments in which they operate exhibit many of the properties characteristic of complex systems, including nonlinear growth patterns, emergent phenomena, and cascading effects that can lead to tail risks. Complexity theory can help illuminate the features of AI that pose central challenges for policymakers, such as feedback loops induced by training AI models on synthetic data and the interconnectedness between AI systems and critical infrastructure. Drawing on insights from other domains shaped by complex systems, including public health and climate change, we examine how efforts to govern AI are marked by deep uncertainty. To contend with this challenge, we propose a set of complexity-compatible principles concerning the timing and structure of AI governance, and the risk thresholds that should trigger regulatory intervention.