Closing the Algorithmic Black Box: Breakdowns and Patching Strategies in a Public Service Media

Abstract

T

This article examines the organizational trajectory of a news recommender system developed within RTBF, Belgium’s public service media. Based on thirteen months of ethnographic fieldwork, it conceptualizes the algorithm as a black box in the making and investigates how breakdown–repair cycles shaped its embedding and eventual stabilization within the organization. The study identifies two major breakdowns and demonstrates that the subsequent repairs functioned less as transformative solutions than as patching strategies: targeted adjustments that resolved immediate issues while simultaneously reinforcing the system’s legitimacy. By foregrounding these patching strategies, the article contributes to Science and Technology Studies (STS) by extending the literature on breakdowns and repair. It shows that such practices not only address technical vulnerabilities but also reconfigure organizational relations, contain dissent, and gradually contribute to the closure of the algorithmic black box.