The banality of « reasonable suspicions » and their effects: politics and digitisation

Contrary to what one might think, there is no single form of security (national or global) but forms of security that are contradictory and can be mutually destructive. This is the case between predictive security and security based on the penal order. Predictive security is rooted in suspicions and anticipations of behaviour that want to prevent crime and anticipate actions on the basis of risk profiles and statistical categories. This orientation towards intelligence led policing, anticipation of future criminal activities, preemptive and proactive works by border guards, police, intelligence services is not a new phenomenon. It has his roots in the argument of « protecting the society » against risks even if it involves to have not full evidence, bur a certain amount of « reasonable suspicions » that groups (especially travellers) are potentially a danger. The digitization of data has accelerated this trend and has served as a justification for this departure from evidence and presumption of innocence of each individuals with the characterisations of criteria of dangerosity and creations of specific statistical data organising categories and selectors by way of agregation of data and sorting out of profiles of people sharing characteristics with some other individuals which were condemned previously. What does mean this dangerosity by « association » ? How are organised these techniques which are based on the capacity of tracing the participation into past events via the data individuals left on the digital realm, on the correlations they have between them even if no causality may be discovered, on the anticipations of futures risks by the profilers, the algocrats or in a more banal way by the systems engineers who have built basic algorithms?