Nick Seaver, “Interpretability, or Learning to Listen to Algorithms” (Leading Voices Speaker Series)

How do algorithms work? As algorithmic systems—from Google’s search engine to Facebook’s newsfeed to Spotify’s music recommender—have become objects of popular concern, this question has proven vexing. Not only are these black boxes hidden from public view and illegible to the untrained eye, they are also complex, distributed systems. With the advent of techniques like deep learning, algorithmic systems are often described as “uninterpretable”—so complex that it is impossible, even for insider experts, to explain their outputs. And yet, engineers, like ordinary users, are tenacious interpreters, eager to make sense of algorithmic behavior, regardless of its internal complexity. In this talk, I draw on ethnographic fieldwork with developers of algorithmic music recommenders in the US to describe how engineers interpret supposedly uninterpretable systems. Ironically, “interpretable” systems require very little interpretive work, while “uninterpretable” systems occasion much interpretation. For music recommenders, this interpretive work is often literally or figuratively akin to listening—an aural, informal sensemaking practice that plays a key role in the nominally rational process of system building and maintenance. Learning to listen to algorithms offers a way to understand how these increasingly powerful technical systems come to work as they do.