What question did this study set out to answer?

The research examines the limitations of Rawlsian fairness in the context of evolving artificial intelligence, particularly in societal decision-making.

March 14, 2026Fudan Journal of the Humanities and Social Sciences0 citationsOpen Access

Beyond the Veil: why Rawlsian Fairness is Too Blind

DWDwayne WoodsPurdue University West Lafayette

Key Points

The research examines the limitations of Rawlsian fairness in the context of evolving artificial intelligence, particularly in societal decision-making.
Analyzed the application of Rawls' theory to AI systems
Conducted a case study in healthcare resource allocation
Identified systemic biases and feedback loops in algorithms
Demonstrated that algorithms meeting Rawlsian fairness can still harm vulnerable groups
Highlighted the paradox of unchanging justice principles in dynamic systems
Illustrated how feedback loops can exacerbate existing disparities

Abstract

Artificial intelligence systems increasingly influence critical societal decisions, from healthcare resource distribution to criminal justice evaluations. While John Rawls’ theory of justice, especially his “veil of ignorance,” has become a prominent framework for promoting algorithmic fairness, this paper reveals a fundamental paradox in its use: unchanging principles of justice, when applied in evolving AI systems, can sustain or worsen existing inequalities. We demonstrate how algorithmic systems that meet initial Rawlsian fairness standards can create growing disparities through feedback loops and systemic biases. Our healthcare case study shows how seemingly neutral resource allocation algorithms can systematically harm vulnerable groups even when they meet formal fairness criteria.

Perguntar à IA

Bookmark

View Full Paper