A Discussion of Harm Reduction for Algorithmic Fairness: Lessons from the Opioid Crisis

Abstract

an extension and modification of recent work by Altman et al. which integrates a harm reduction framework into algorithmic design and implementation, but by failing to integrate the voices of those adversely impacted by algorithmic design, does not go far enough to protect the vulnerable and marginalized and fails to incorporate the valuable insights of people with lived experience being processed by algorithms. A fuller harm reduction approach to designing algorithmic fairness makes space for situated ethnographic research of users processed by algorithms and involvement of those users in iterative redesigns.