arXiv 2502.00012
Lessons from complexity theory for AI governance
By Noam Kolt, Michal Shur-Ofry, et al.
Published 2025-01-07
Discussion
Read the public discussion and references gathered around this paper.
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 t…