Learning & Games Reading Group: Learning in the Presence of Strategic Behavior

Title: AI, algorithmic pricing and Collusion

Abstract: Increasingly, algorithms are supplanting human decision-makers in

pricing goods and services. To analyse the possible consequences, we

experimentally study the behavior of independent tabular RL algorithms

interacting with each other and maximising profits in a synthetic

marketplace. We find that the algorithms consistently learn to charge

supra-competitive prices, without communicating with one another. The

high prices are sustained by collusive strategies with a finite phase

of punishment followed by a gradual return to cooperation. In the

second part of the talk, I will summarise a policy debate on

algorithmic pricing bridging together academics in Law, Economics and

Computer Science and introduce a field experiment meant to test the

external validity of our findings.