I just read the second piece from the TCS special issue on Codes and Codings in Crisis (see my previous post). It’s a typically brilliant piece by Louise Amoore. It extends her recent work on the way that digital data and software are used in making security decisions. This work has focused in particular on how border guards are helped in assessing the risk posed by individuals.
In this particular article, which covers some difficult issues in more detail than I can cover here, Amoore describes the influential presence of ‘data derivatives’ in decision making. These descriptions focus on how fragmented data resources are used to create derivatives, which in term inform the algorithmic decisions made about risk. Often these derivatives are formed from resources that were not intended for this purpose. They can even be created from an absence of available data which lends itself to uncertainty and risk. There appears then to be an underlying politics in these security systems and Amoore seems to be concerned with unpicking these decision making processes and the data resources they use. Amoore’s piece moves toward an analysis of how data derivatives are used to make predictions or to forecast potential actions. As with the introductory piece by Mackenzie and Vurdubakis, this shows how code has real consequences for people and how wider social forces become real through the work of code.