What is the Point of Fairness? Disability, AI and The Complexity of Justice

Fairness frameworks alone are insufficient for protecting disabled people from AI harms. The paper argues that focusing narrowly on "fairness" in AI ethics can actually reinforce existing power imbalances and medical gatekeeping structures, rather than addressing the root causes of injustice that disabled people face.

AI systems designed to "help" disabled people can cause harm to the most marginalized. Through case studies in computer vision technology, the authors demonstrate that tools intended to assist disabled people may benefit relatively privileged disabled individuals while causing harm to those who face multiple forms of marginalization—a problem that fairness metrics alone fail to capture.

Structural injustices must be considered during AI design, not just evaluated after deployment. The paper emphasizes that ethical AI development for disabled people requires examining how broader systems of inequality shape technology from the beginning, rather than treating fairness as a technical problem to be solved through algorithms.

Researchers should shift from "fairness" to "justice" as their ethical framework. The authors call for AI ethics and disability scholars to adopt justice-centered approaches that account for complex, intersecting inequalities rather than relying on simplified fairness concepts that may obscure whose interests are actually being served.