"With the aid of your precog mutants, you've boldly and successfully abolished the post-crime punitive system of jails and fines. As we all realize, punishment was never much of a deterrent, and could scarcely have afforded comfort to a victim already dead."

Philip K. Dick, The Minority Report

In his short story ‘The Minority Report’, science fiction author Philip K Dick has one of his main characters suggest that a pre-emptive strike of precrime ‘punishment’ should comfort the victim, claiming that the life of the victim is more important than an individual human subject’s ability to develop her moral agency. This pointedly confronts us with the dilemma of the criminalization of future behaviours. If criminal intelligence is capable of predicting who will most probably commit a murder, should we punish the culprit before she can perform the act? The reader may object that in such a case the suspect cannot be punished, as punishment cannot refer to future action; as long as a person has not performed the action her behaviour cannot be qualified as wrongful and thus cannot be liable to punitive intervention. The right term would be something like a preventive measure, such as detention, therapy or any kind of physical intervention that would rule out the criminal act that has been predicted. I invite the reader, however, to imagine that having been categorized as a person that will commit a crime at some point in the future could indeed lead to an accusation of wrongfulness and culpability. Contrary to our present common sense, this wrongfulness would inhere in the fact that a person will violate the criminal law, and this future violation would also imply guilt. That we now think this to be nonsensical is no guarantee that epistemic changes triggered by novel computing infrastructures could not instigate an entirely different concept of what it means to punish a person.

In this chapter I will explore the implications of this thought-experiment, suggesting that proactive forensic profiling may extend the boundaries of the criminal law in a way that surreptitiously erodes the very meaning of punishment as we understand it today. Such an exercise may sharpen our awareness of the achievements of the legal protection presently offered by the criminal law and confront us with the extent to which we take such protection for granted. Building on findings within the field of profiling practices I will argue that new ways of knowledge construction will challenge the logic of the criminal law. First, they may counter our expectation that the criminal law responds to past rather than future events, and second, they may counter our expectation that the criminal law concerns actions rather than biological or behavioural characteristics. Those who prefer to think of the central tenets of criminal law as moral or conventional maxims that are independent of socio-technical infrastructure, could find my argument a tiring exercise. They may not be interested in the technicalities of what has been called ‘knowledge discovery in databases’ (KDD). However, if as has been argued by a number of legal scholars the epistemic shift generated by the digital revolution has serious implications for the meaning of law, this would also concern the meaning of punishment. To understand these implications we need to come to terms with the technologies that trigger them. This chapter thus offers a tentative exploration of the implications of emerging socio-technical infrastructures of knowledge production for the scope of the criminal law. After a brief discussion of the meaning of terms like criminal and forensic profiling, I undertake an analysis of proactive forensic profiling as producing novel types of knowledge claims that are highly relevant for proactive criminalization. I will argue that these knowledge claims, typical of actuarial justice, are an affordance of the socio-technical infrastructure of profiling technologies. The novelty of ‘knowledge discovery in databases’ concerns its focus on effective prediction without the need to understand or explain the patterns it uncovers. This brings in the issue of human autonomy in relation to causal determination and freedom of the will: what does it mean that profiling technologies are capable of ‘predicting’ our future behaviours? Rejecting both determinacy and indeterminacy as problematic Cartesian viewpoints, I embrace the underdeterminacy of human agents as situated embodied subjects, and investigate how such underdeterminacy relates to profiling and to the boundaries of criminalization in the present legal framework. In the last section I will argue that the knowledge claims generated by profiling techniques can restrict human freedom if they profile people as correlated objects, whereas they can also enlarge human freedom if they allow people to become aware of the profiles they match. People can realise their potential as correlatable subjects, capable of resisting the correlations they are presented with. Returning to the main point of this chapter, I will conclude that the issue of whether human agents are treated as correlated objects or as correlatable subjects is not just a matter of moral philosophy, but a matter of engaging in the design of the architecture of proactive forensic profiling.