1171 | 205 | What AI Does and Does not See: De-scribing Intelligent Pedestrian Traffic Lights in Vienna

In this paper, I investigate a mundane urban technological intervention powered by AI in Vienna. The intelligent pedestrian traffic lights initiative (IPTL) employs surveillance cameras and algorithmic processors to recognize pedestrians’ intention to cross a crosswalk in order to change the signal for them automatically. The aim is to prevent jaywalking and provide a better flow of traffic. My investigation denotes the ways in which the experts define, design and test AI in the city throughout the development of the initiative. I ground my analysis in a historical perspective on the changing roles of traffic lights in cities. Here, the IPTL opens a space to revisit the invisible infrastructure of urban temporal ordering. Throughout the analysis, I draw attention to the co-constitution of (in)visibilities. I show how the initiative is locked-in with the experts’ tech-solutionist imagination where the cycle of problem-solution lacks a sociotechnical understanding of streets and mobility. Thus, I identify conflicting imaginaries of use and users. In the end, I propose a reconstructive approach to rethink how the initiative could be otherwise by being attentive to the co-constitution of (in)visibilities in urban technological interventions especially at the advent of AI in the city.

Pouya Sepehr
University of Vienna; University of Turin

ID Abstract: 205