Regulating eXplainable Artificial Intelligence (XAI) May Harm Consumers

TL;DR

This paper uses a game theoretic model to show that regulating AI for full transparency through XAI can sometimes harm consumers and reduce overall social welfare, challenging common assumptions.

Contribution

It introduces a game theoretic framework analyzing XAI regulation effects and highlights potential negative impacts on welfare and the concept of XAI fairness.

Findings

XAI regulation may be redundant in some cases.

Mandating full XAI transparency can worsen outcomes for firms and consumers.

XAI fairness may be impossible to guarantee even with regulation.

Abstract

Recent AI algorithms are black box models whose decisions are difficult to interpret. eXplainable AI (XAI) is a class of methods that seek to address lack of AI interpretability and trust by explaining to customers their AI decisions. The common wisdom is that regulating AI by mandating fully transparent XAI leads to greater social welfare. Our paper challenges this notion through a game theoretic model of a policy-maker who maximizes social welfare, firms in a duopoly competition that maximize profits, and heterogenous consumers. The results show that XAI regulation may be redundant. In fact, mandating fully transparent XAI may make firms and consumers worse off. This reveals a tradeoff between maximizing welfare and receiving explainable AI outputs. We extend the existing literature on method and substantive fronts, and we introduce and study the notion of XAI fairness, which may be impossible to guarantee even under mandatory XAI. Finally, the regulatory and managerial implications of our results for policy-makers and businesses are discussed, respectively.