Sociology of Algorithms
This workshop aims to bring together scholars and practitioners interested in analyzing the social embeddedness of algorithms and intervening in the processes through which their results are interpreted and used for a variety of social purposes. Complex algorithms are increasingly called upon as a purportedly objective substitute for and/or enhancement of expert judgment in a multitude of tasks ranging from predicting dangerousness of potential parolees, through probabilistic DNA profiling (PDP) used to establish crime scene DNA matches, to determining welfare eligibility and flagging potential fraud. The workshop seeks to explore five dimensions of a nascent sociology of algorithms: 1) critical analysis of how algorithms are leveraged to automate systems of social control and surveillance, and with what consequences for already marginalized communities; 2) inquiries that are calculated to shed light on the hidden work that various social actors do to fit algorithms – by selecting and digitizing their input and repairing their output - within collectives made of humans and machines; 3) exploring the myriad dimensions of data provenance, such as consent, privacy, exploitation, partnership/sharing, as well as data colonialism, the fallibility of proxy variables, etc.; 4) comparative analysis of the jurisdictional struggles provoked by the introduction and increasing reliance on algorithms in multiple fields to perform tasks previously performed by experts; 5) attention to the representational struggles, or “trials of attribution,” around who can speak for the algorithm or in what form is the algorithm made to speak. The workshop seeks to combine approaches informed by Science and Technology Studies (STS) and the sociology of expertise, with critical approaches that are attuned to the racial, gender and class inequalities and biases built into algorithms and reproduced by their functioning. Finally, a unique aspect of the workshop is bringing together academics with practitioners who are attempting to shape or limit the influence of algorithms in their own field of work (e.g. defense lawyers). We envision the workshop as an incubator for collaborations across the boundary between academia and practitioners “in the wild.”