Algorithmic Auditing and Social Justice: Lessons from the History of Audit Studies

Abstract

Algorithmic audits" have been embraced as tools to investigate the functioning and consequences of sociotechnical systems. Though the term is used somewhat loosely in the algorithmic context and encompasses a variety of methods, it maintains a close connection to audit studies in the social sciences-which have, for decades, used experimental methods to measure the prevalence of discrimination across domains like housing and employment. In the social sciences, audit studies originated in a strong tradition of social justice and participatory action, often involving collaboration between researchers and communities; but scholars have argued that, over time, social science audits have become somewhat distanced from these original goals and priorities. We draw from this history in order to highlight difficult tensions that have shaped the development of social science audits, and to assess their implications in the context of algorithmic auditing. In doing so, we put forth considerations to assist in the development of robust and engaged assessments of sociotechnical systems that draw from auditing's roots in racial equity and social justice.

A Silicon Valley love triangle: Hiring algorithms, pseudo-science, and the quest for auditability

“…In addition to legal and data science teams that are currently engaging in algorithmic audit work, we add our voices to the chorus of calls for addition of social scientists, psychologists, and historians of science and technology to critically evaluate assumptions and epistemologies and inform the audit process as a whole. 30,[53][54][55][56][57]…”

Fair ranking: a critical review, challenges, and future directions

“…There are two broad lessons from this literature, that we explain and apply to the design of fair RS, in a manner that involves integrated effort from different actors and a comprehensive view of their effects. First, as discussed by Vecchione et al [155], a key point when assessing or auditing algorithmic systems is to move beyond discrete moments of decision making, i.e., to understand how those decision-points affect the long-run system evolution; this point is particularly true for fairness interventions in ranking and recommender systems, as discussed in Section 3. Jannach and Bauer [81] also highlight the limitations and unsuitability of traditional research in RS, which focused solely on accurately predicting user ratings for items ("leaderboard chasing") or optimizing clickthrough rates.…”

Accountability in an Algorithmic Society: Relationality, Responsibility, and Robustness in Machine Learning

“…In certain cases, scholars have re-framed accountability standards around harmed and vulnerable parties [93,110]. This work -particularly that which focuses on transparency [39] and audits [135] -makes clear that standards of care and frameworks, while important for developing actionable notions of accountability, do not guarantee accountability on their own. Algorithmic impact assessments (AIAs) attempt to fill this gap [96].…”