Cloud Ethics

This ambitious work is a rich and complex response to the ascent of machinelearning algorithms. More specifically, it is an effort to examine the philosophically sophisticated nature of this rising power. It hardly needs to be said that the work is timely, as computational algorithms occupy increasingly central positions not only in the economies of production and distribution but also in the economies of information, a function that extends to the management of social life in general and to the exercise of state security, surveillance, and policing powers in particular. Amoore’s exposition begins, in fact, with a quietly harrowing account of the partnership between Geofeedia, a ‘location-based analytics platform,’ and the Baltimore Police Department during the period following the death of Freddie Gray, a young black man who suffered fatal injuries while in police custody in 2015. In the civil unrest that followed, ‘terabytes of images, video, audio, text, and biometric and geospatial data from the protests of the people of Baltimore were rendered as inputs to the deep learning algorithms’ (p. 3). Many protesters were arrested or detained without charges based solely on the presumed authority of the predictive algorithm, which had ‘learned how to recognize what a protest is, what a gathering of people in the city might mean’ (pp. 3–4). The book is organized in three main sections, preceded by an introduction in which the author sets out her approach to the ethicopolitics of algorithms and distinguishes it from more familiar public calls for vigilance: calls to divest algorithms of their racial biases, for example, or to make them more transparent so that those responsible for their creation can be held accountable. While Amoore recognizes the palpable threats against persons and against rights identified in these calls for action, she seeks to identify effects working at a more fundamental and esoteric level. ‘In short,’ she writes, ‘what matters is not primarily the identification and regulation of algorithmic wrongs, but more significantly how algorithms are implicated in new regimes of verification, new forms of identifying a wrong or of truth telling in the world’ (pp. 5–6).