My dissertation, The Algorithm at Work: The Reconfiguration of Work and Expertise in the Making of Similarity in Art Data, was an ethnography of a start-up firm commonly referred to as the 'Pandora for Art.' Organized around a similarity matching algorithm and a large, continuously changing database of art images, the technology provided an expert-oriented framework for end-users to discover their tastes and enter the emerging online market for contemporary art.
I studied the work practices of a team of art experts, who were hired to make art image data legible to the algorithm and make the algorithm’s output legible to those with knowledge about art. Observing in-person and online interactions in moments when the technology broke down allowed me to see the contention and negotiation involved in the ongoing repair of the algorithm by the art experts. Despite limited technical skills, the team was continuously mobilizing competing claims about how the algorithm worked in order to repair breakdowns between the algorithm's output and their expectations. As they experimented with different explanations for these breakdowns, a novel form of expertise emerged that created opportunities for collaboration and conflict on the team.
My research contributes to existing literatures on standardization, expertise, repair work, and teams, as well as to the emerging literature on the sociology of data science. More importantly, the study highlights the undervalued skills increasingly required by the expanding market of contingent knowledge labor associated with data analytic technologies.
Sachs, S.E. 2020. “The algorithm at work? Explanation and repair in the enactment of similarity in art data.” Information, Communication, & Society.