Aleš Završnik

The transformation of an increasingly greater part of our activities into a digital language creates an enormous quantity of (personal) data – "big data" – on which the existential functioning of postmodern society depends. The existence of "big data" provides many advantages to individuals, companies, and states, as social networks suggest whom to befriend, algorithms trade our stocks, and computers help the military find its targets. But at the same time big data is a source of numerous difficulties. Algorithms may incorporate biases of their creators. They may circumvent prohibitions of the use of sensitive personal data in judicial proceedings, such as religion or ethnic origin, by calculating these attributes from other indirectly related data, such as time of consummation of a particular food. There are other severe caveats of algorithmic decision-making, such as shifting responsibility as the accountability for the final decisions would lie not only in the hands of, for instance the judge sitting in court, but also in the hands of the remote programmers and consultants designing the code.

In modern western societies the predominant belief is that digital technologies and big data will solve numerous social problems, from the economic crisis to terrorism. This "revolution that will transform how we live, work, and think" (Mayer-Schönberger & Cukier 2014) is shifting power relations in the so-called "security and control" domain as well. Criminal justice systems are using technological solutions too, for instance to predict the future crimes of those applying for bail or those to be released on parole. Predictive policing software, such as IBM's Blue Crush program, produces probability reports on criminality and assures us that societies will reduce crime by using it. The USA and the EU are financing research on the development of computer programs that are able to predict criminal behaviour on the basis of recognising facial or "walking" characteristics. Predictive policing software, such as the TrapWire program, looks for patterns that would help predict terrorist attacks. Predictions are made by using data input from CCTVs, sometimes installed all across the city; facial recognition data from CCTV networks are then combined with data from various databases.

Such knowledge, built on a large amount of seemingly unrelated data, whose credibility is based on complex mathematical algorithms, legitimises increased social control, limits privacy and undermines the basic principles of criminal procedure. Underlying these and many other potential uses of big data in crime control, however, are a series of legal and ethical challenges relating to, inter alia, privacy, discrimination, and the presumption of innocence.