Art of Crowd Control: The Psychic Life of the Masses in Computer Vision

• In Event: Total Institutions: Technosciences of Care and Control III

Abstract

In the era of Black Lives Matter, crowds of protestors, whether viewed from above or below, have formed a new “surface” for computer vision AI. There is little transparency on police theoretical knowledge of crowds and applied practice, and yet, the unprecedented state imaging and digitization of the crowd continues to mount, producing a unique set of questions and “algorithmic anxieties” distinct in this new perception of “crowd” at the intersections of crowd psychology, art, technology, and policing. This paper examines the coupling of crowd analytics research with protest surveillance which has inaugurated the imaging of the “masses” with new algorithmic detections: from predicting a protest’s tendencies towards violence to identifying deviant political affiliation. The promise of crowd analytics lies in scale, presumably exceeding human vision by containing the crowd into manageable algorithms. However, I argue crowd analytics research as it currently trends replicate a long-critiqued classic urban theory that essentializes and racializes crowd as unconstrained and dangerous, one shared by the contentious history of urban policing. I conclude with a reading of the Bell Labs artistic collaboration by media artist Sougwen Chung and video analytics researcher, Larry O’Gorman to consider the “ethico-political” potential of crowd analytics algorithms (Amoore 2020). Their collaboration explores the computer vision of crowds through the potential of expressivity and emotionality, rather than racist pathology and policing.

Author

• Peggy Lee, Emory University