Surfacing Algorithms: An Inventive Method for Accountability

Abstract

For many scholars and policy makers, the democratic demand for algorithmic transparency is a call for openness. Yet critical algorithm studies have shown that classical accountability devices, such as freedom of information, audits, or code openness have failed to make algorithms meaningfully knowable to ordinary citizens. Rather than waiting for curated and limited transparency to open so-called “hidden black boxes,” I propose and illustrate in this article a method of “surfacing algorithms.” This experimental method makes algorithms accountable through the (re)design of their surfaces: the forms of appearance, documentation, and tangible devices accompanying their use in everyday life. Democratic accountability is thus reconceptualized as a design problem. I demonstrate my method by showing how the reverse-engineering of the French housing tax algorithm performed by citizens during collective workshops through the scrutiny of their tax letters served to interrogate a fiscal algorithm from its surface.