Human 'irrationality' is the only thing preventing AI pricing algorithms from colluding to hike prices.
In a simulated market, pricing bots found it easy to collude and overcharge 'rational' consumers. However, when they faced humans with psychological 'biases' like reference dependence (caring about past prices), the algorithms were forced to compete harder, paradoxically making the 'irrational' behavior a safeguard for market welfare.
Consumer Behavior and Algorithmic Price Competition: The Case of Reference Dependence
SSRN · 6388018
As AI and humans increasingly coexist in the marketplace, it is important to understand how human psychology interacts with algorithmic decision-making. How do consumer behavioral patterns influence algorithmic pricing and the resulting profits and consumer welfare? Should consumer AI agents emulate human psychology or act rationally? While current wisdom suggests removing human biases to maximize efficiency, our paper investigates how consumers' reference dependence, a core human "bias", shapes