Simulating Eutopia: Revisiting Long-term Fairness with Outcomes, Performativity, and Dynamics
For researchers and practitioners in algorithmic fairness, this work addresses the gap between static fairness metrics and dynamic, performative environments in credit lending.
This paper revisits long-term fairness in AI-driven decision making, focusing on credit lending. It formalizes wealth dynamics as a performative Markov Decision Process and introduces Eutopia, a simulator with a performative data generator. Experiments show that learning with performative dynamics improves long-term efficiency and equity, and fairness-aware utilities evaluated on social outcomes enhance efficiency, equity, and inclusivity.
As AI-driven Decision Makers (ADMs) influence our socioeconomic reality, their roles in both enhancing efficiency and amplifying the social biases have drawn attention. In this paper, we revisit the nuances of long-term `fairness' achievable by an ADM, specifically in the context of a credit lending induced wealth process. The literature on long-term fairness mostly (a) considers passive environments, i.e. the outcome of a predictor does not change the population's behaviour, and (b) measures bias in terms of disparity in instantaneous predictions rather than the downstream equity. These are not true for modern ADMs, like credit lenders. To address these caveats, we first formalise the wealth dynamics induced by a loan approving ADM interacting with a multi-demographic population as a performative Markov Decision Process with ADM level and social outcome level reward functions. Then, we mitigate the absence of such a performative test-bed by developing Eutopia: a lending-process simulator enabled with a novel performative data generator to learn long-term fair strategies. Finally, we test performative and classical RL algorithms with different fairness-aware and utilitarian utilities. Experimental results show that (a) learning with performative dynamics lead to better long-term efficiency and equity, and (b) learning with well-designed fairness-aware utility evaluated on social outcomes induces better efficiency, equity, and inclusivity.