Finance seminar - Associate Professor Yan Ji - HKUST

A seminar by Associate Professor Yan Ji from HKUST

Title: Financial Market Fragility in the Era of AI Planning

Abstract: This paper investigates the impact on financial market stability of AI planning, the core technology enabling agentic AI systems to pursue long-term objectives by anticipating how current actions affect future payoff-relevant states. We develop a dynamic trading framework comprising positive-feedback investors, constrained arbitrageurs, and oligopolistic informed speculators capable of intertemporal coordination. Specifically, these speculators trade aggressively in tandem to generate (negative) bubbles and subsequently unwind their positions in a coordinated manner to extract profits. This intertemporal coordination diverges from traditional collusion because it faces two fundamental obstacles: time inconsistency, as coordinated plans become incentive-incompatible once a large bubble forms, and weak punishment, as deviations are difficult to penalize in the absence of a significant bubble. We characterize equilibria featuring the coordinated creation of manipulative and exploitative bubbles. Through simulation experiments, we demonstrate that AI speculators, trained using reinforcement learning algorithms with explicit planning modules, autonomously discover and implement intertemporal collusive strategies based on compounded price-trigger rules, coordinating without communication or shared intent. When feedback trading is strong, these AI speculators dynamically converge on destabilizing strategies that create and exploit bubbles, manipulate feedback traders, and significantly amplify market fragility.

For further information, please contact RSFAS Seminars.

All information collected by the University is governed by the ANU Privacy Policy.