[2603.20281] On the Fragility of AI Agent Collusion

Computer Science > Computer Science and Game Theory arXiv:2603.20281 (cs) [Submitted on 18 Mar 2026] Title:On the Fragility of AI Agent Collusion Authors:Jussi Keppo, Yuze Li, Gerry Tsoukalas, Nuo Yuan View a PDF of the paper titled On the Fragility of AI Agent Collusion, by Jussi Keppo and 3 other authors View PDF HTML (experimental) Abstract:Recent work shows that pricing with symmetric LLM agents leads to algorithmic collusion. We show that collusion is fragile under the heterogeneity typical of real deployments. In a stylized repeated-pricing model, heterogeneity in patience or data access reduces the set of collusive equilibria. Experiments with open-source LLM agents (totaling over 2,000 compute hours) align with these predictions: patience heterogeneity reduces price lift from 22% to 10% above competitive levels; asymmetric data access, to 7%. Increasing the number of competing LLMs breaks up collusion; so does cross-algorithm heterogeneity, that is, setting LLMs against Q-learning agents. But model-size differences (e.g., 32B vs. 14B weights) do not; they generate leader-follower dynamics that stabilize collusion. We discuss antitrust implications, such as enforcement actions restricting data-sharing and policies promoting algorithmic diversity. Comments: Subjects: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI) MSC classes: 91A20, 91B24, 68T05 ACM classes: J.4; I.2.11; K.4.1 Cite as: arXiv:2603.20281 [cs.GT]   (or arXiv:2603.20281v1 [cs.G...