Algorithmic Fairness in the Roberts Court Era

Abstract

Scientists and policymakers alike have increasingly been interested in exploring ways to advance algorithmic fairness, recognizing not only the potential utility of algorithms in biomedical and digital health contexts but also that the unique challenges that algorithms—in a datafied culture such as the United States—pose for civil rights (including, but not limited to, privacy and nondiscrimination). In addition to the technical complexities, separation of powers issues are making the task even more daunting for policymakers—issues that might seem obscure to many scientists and technologists. While administrative agencies (such as the Federal Trade Commission) and legislators have been working to advance algorithmic fairness (in large part through comprehensive data privacy reform), recent judicial activism by the Roberts Court threaten to undermine those efforts. Scientists need to understand these legal developments so they can take appropriate action when contributing to a biomedical data ecosystem and designing, deploying, and maintaining algorithms for digital health. Here I highlight some of the recent actions taken by policymakers. I then review three recent Supreme Court cases (and foreshadow a fourth case) that illustrate the radical power grab by the Roberts Court, explaining for scientists how these drastic shifts in law will frustrate governmental approaches to algorithmic fairness and necessitate increased reliance by scientists on self-governance strategies to promote responsible and ethical practices.