From selective to total social control through data science systems?

• In Event: Data acquiring, sharing and analysing for better Criminal Policy Making

Abstract

Selectivity appears to be the key operating principle of social control systems. Formal as well as informal social control is successful because it intervenes only in selected quantities and qualities of deviant and delinquent behaviour. The filtering process of law enforcement and strategies of non-intervention or diversion reflect this mechanism. It has evolved under con-ditions of limited information and limited possibilities of data analysis. Today, however, there seems to be the technical possibility of collecting all information on deviant behaviour and deviant individuals (big data) and analysing it by means of machine learning. Such an infor-mation management would open up possibilities for total control and intervention strategies, not only ex post, but also and especially ex ante. The extent to which this might be feasible will be discussed, taking into account the potential and limitations of machine learning sys-tems in areas such as social scoring or predictive policing. And to what extent legal re-strictions, such as the EU AI Act, will be reliable to resist the "seduction of totality". The main assumption of the talk is that for a successful social control system, total surveillance cannot replace the principle of selectivity.

Authors

• Klaus Boers, University of Muenster

• Marcus Schaerff, University of Münster