Incentive Aware AI Regulations: A Credal Characterisation
This work provides a foundational framework for developing enforceable AI regulations, particularly for high-stakes ML applications, by connecting mechanism design and imprecise probability.
This paper addresses the problem of strategic ML providers evading regulations by framing AI regulation as a mechanism design problem under uncertainty. The authors introduce regulation mechanisms that map empirical evidence to market share licenses, forcing providers to bet on their model's compliance. They prove that a mechanism achieves a perfect market outcome (non-compliant providers self-exclude, compliant providers participate) if and only if the set of non-compliant distributions forms a credal set.
While high-stakes ML applications demand strict regulations, strategic ML providers often evade them to lower development costs. To address this challenge, we cast AI regulation as a mechanism design problem under uncertainty and introduce regulation mechanisms: a framework that maps empirical evidence from models to a license for some market share. The providers can select from a set of licenses, effectively forcing them to bet on their model's ability to fulfil regulation. We aim at regulation mechanisms that achieve perfect market outcome, i.e. (a) drive non-compliant providers to self-exclude, and (b) ensure participation from compliant providers. We prove that a mechanism has perfect market outcome if and only if the set of non-compliant distributions forms a credal set, i.e., a closed, convex set of probability measures. This result connects mechanism design and imprecise probability by establishing a duality between regulation mechanisms and the set of non-compliant distributions. We also demonstrate these mechanisms in practice via experiments on regulating use of spurious features for prediction and fairness. Our framework provides new insights at the intersection of mechanism design and imprecise probability, offering a foundation for development of enforceable AI regulations.