This text explores how the introduction of (semi)autonomous cars fundamentally transforms traffic accidents. Using examples of recent crashes, the article examines how edge cases, accidents, and machine learning are intertwined and produce “infrastructural violence.” Following the example of the suspension of Cruise in 2023, the text investigates edge cases, which may not coincide with actual collisions but necessarily raise their probability, because they enable constant optimization through machine learning. I argue that it is therefore necessary for mobility studies to go beyond existing frameworks in order to investigate the epistemology of accidents of self-driving cars and to scrutinize the transitions from human perception to big data, machine learning, and sensor-based world modeling.
Florian Sprenger is professor for virtual humanities at the Department of Media Studies at Ruhr-Universität Bochum. His research covers the transformations of digital cultures, the history of artificial environments, and the virtuality of machinic worldmaking. Email: florian.sprenger@rub.de