Accidents Happen: Three Evaluations of Algorithmic Accidents

• In Event: Transforming Subjects into Objects

Abstract

Accidents happen, both in the non-digital world and in the statistical/algorithmic. In this paper, I will explore the concept of accident, focusing specifically on Catherine Malabou’s interpretation of the accident as a moment of emergence, a “new being” that disrupts our sense of historicizing and discourse and thus facilitates new knowledges of the world. Traditionally, these accidents are understood as aberrations from the norm, or disruptions. I will argue in opposition, proposing a new theory of the accident for our algorithmic times, one that gestures towards how statistical/algorithmic accidents produce emergence/new knowledges not through difference or disruption, but through sameness. This argument comes from a tour across three examples, from statistical analysis disproving the authenticity of an aristocrat’s signature due to its deemed-impossible similarity to another, to essentialist, and thus homogenizing, statistical analyses used to prove the guilt of an interracial couple, and to the physical accident of an algorithmically-driven car the occurred because sensors were unable to differentiate the white sky from the white truck turning into its lane.

Author

• John Cheney-Lippold, University of Michigan