Topic: AI contagion in social networks

Lecturer: Prof. Olivier BOS

Time: March 5th, 2026, 14:30-15:30, UTC+8

Venue: Meeting Room 311, School of Management

Biography: Professor Olivier BOS is the Deputy Head of the Social Sciences Department of ENS Paris-Saclay, the Director of the Centre for Economics at Paris-Saclay, and a doctoral supervisor. He graduated from the Paris School of Economics and was awarded the title of Junior Member at the Institut Universitaire de France (IUF). His main research interests lie in game theory and industrial organization, and he has published numerous academic papers in international journals. More recently, he’s turned his attention to political economy and the application of AI in economics and management.

Abstract: We’ve investigated how AI interacts with social communication networks to affect the stability of collective knowledge. Agents exchange information through social networks and simultaneously receive content generated by AI. The AI system is retrained on the aggregate information of the society it influences. This interaction creates two feedback channels. The first is an AI contagion channel that spreads distorted information within the network. The second is an AI social distortion multiplier, through which AI retraining amplifies past errors. Despite the environment’s high dimensionality, we demonstrate that the system’s long-term behaviors are two-dimensional on a stable trajectory, and that the spectral radius determines whether the AI-mediated information system is stable or dynamically unstable. We characterize a clear regulatory frontier that sets the minimum filtering level required to achieve stability and show how network topology shapes systemic information risks.

Rewritten by: Mei Mengqi

Edited by: Li Huihui, Li Tiantian

Source: School of Management