Využití algoritmů při profilování v trestním řízení a důsledky pro lidská práva
The Use of Algorithms for Profiling in Criminal Proceedings and Implications for Human Rights

Summary/Abstract: Algorithms are replacing activities that have previously been performed by humans. One of these activities is judicial decision making based on algorithms. It is generally assumed that algorithmic decision making, as opposed to the human decision making, is free from prejudice and bias. This assumption is problematic in two ways. Firstly, the algorithm decides according to pre-specified criteria and values. These values are chosen by humans who may include among the predefined criteria the criteria that correspond to their prejudices and preconceptions. Secondly, some algorithms may have their own learning ability, which means that the generated results are adapted to the previous ones to which were given preference in the past. The users may have the possibility to choose precedents that correspond with their point of view, which may be based on prejudice. In the United States and also in some European countries an algorithm is already being used by courts to assess whether or not it is likely that the accused (or condemned) will commit another crime in the future. The judge does not decide based on the circumstances of the case and individual characteristics of the accused. The judge decides instead on the basis of a profile. The profiles are based on personal data of the offender. Some of those data have a character of biometric data. The algorithmic decision making based on profiling raises concerns about the rights of the accused in the judicial proceedings, namely the prohibition of discrimination and the right to a fair trial.

• Details
• Contents