Against Access

• In Event: Black Boxes at Work: Approaches to Studying Corporate Software in the Wild

Abstract

As algorithms have gained notoriety among humanists, social scientists, and the general public, their defining feature has shifted from computation to secrecy. Algorithms may still be rationalizing, quantifying, and procedural, but first they are secret—hidden behind corporate legal protections or within complex computational systems. The blackness of the black box has gone from signifying a lack of concern for interiors to signifying hiddenness. In this situation, the social study of algorithms has been deformed, turned from an interest in the politics and pragmatics of computation to the acquisition of corporate secrets—piercing the legal shields constructed by corporations, or, to use a common metaphor that demonstrates the colonial and gendered politics of this approach, "opening the kimono."

In this paper, I argue against the privileging of access in the production of knowledge about algorithms. I make this argument as someone who achieved "access" in a conventional sense, in the course of several years of multi-sited fieldwork with the US-based developers of music recommender systems. Fetishizing access turns ethnographers into couriers of objective facts, rather than interpretations, and it obscures the ordinary opacities of social life, treating secrecy as an obstacle to be heroically overcome, rather than an everyday accomplishment that is continuously entangled with the production of algorithmic systems at all levels. I propose that empirical engagement with algorithmic systems remains a crucial and neglected practice, but that engagement is not circumscribed by something called "access."

Author

• Nick Seaver, Tufts University