Algorithmic Fairness in AI Surrogates for End-of-Life Decision-Making

TL;DR

The paper addresses fairness in AI surrogates for end-of-life decision-making, arguing that traditional, distributive fairness notions fail in existential, relational contexts. It maps core fairness concepts to ethical risks in DNR decisions, and advocates a moral pluralism framework that respects relational autonomy and cultural diversity. A key contribution is a governance blueprint—including ethics translation boards, CARE-driven data stewardship, dynamic consent, and continuous fairness audits—designed to operationalize fairness as moral attunement rather than sole outcome parity. The work emphasizes cross-cultural perspectives, stakeholder engagement, and institutional accountability to ensure AI surrogates support dignity, mitigate epistemic injustice, and augment, not replace, human moral deliberation. Practical impact lies in guiding policy, governance, and design strategies for ethically attuned end-of-life AI systems under high-stakes uncertainty.

Abstract

Artificial intelligence surrogates are systems designed to infer preferences when individuals lose decision-making capacity. Fairness in such systems is a domain that has been insufficiently explored. Traditional algorithmic fairness frameworks are insufficient for contexts where decisions are relational, existential, and culturally diverse. This paper explores an ethical framework for algorithmic fairness in AI surrogates by mapping major fairness notions onto potential real-world end-of-life scenarios. It then examines fairness across moral traditions. The authors argue that fairness in this domain extends beyond parity of outcomes to encompass moral representation, fidelity to the patient's values, relationships, and worldview.