Benefits and Limits of Autonomous Systems in Public Security

Abstract

The replacement of traditional deterministic machines by autonomous, self-organizing systems in the realm of security technologies represents a new challenge. Operative and strategic autonomy must be made more distinct, while moral autonomy should remain a human responsibility. Depending on the type of mechanical learning, the systems gain a black box character that impedes the tracing of the authorship of the systems’ actions. The proposed method of monitoring system decisions (Digital Charta Art. 5, DSGVO Art. 22) must deal with complex networks of responsibility. For this reason, a new legal framework and new political terms of negotiation must be developed.