Abstract
The possibilities and limits of the use of artificial intelligence in law are ever more discussed. The ability to mine, select, and swiftly analyse the data from extensive databases makes artificial intelligence an exquisite instrument for the support of professionals in many fields of law.
However, the hypothetical possibility that artificial intelligence could be directly involved in decision-making either in particular cases (mainly in the judiciary) or in legislation is very controversial for various reasons. Among other technical and value-oriented questions, we need to ask what the implications of the influence of artificial intelligence on these decisions for their democratic legitimacy will be.
For decades, there has been an ongoing debate regarding different types of constitutionalism, especially political constitutionalism which leaves decision-making on political questions to the widest extent to elected legislative bodies, and legal constitutionalism which does not trust the political sphere and justifies more intensive interventions of the judiciary in order to protect basic rights. The core of the dispute lies in the permissible intensity of the judicialization of politics, i.e. deciding political questions by unelected legal experts.
In the case of the important role of artificial intelligence in similar decision-making, it will be necessary to ask to what extent it is appropriate to leave political problems to un-human agents which will use decision-making processes that will be a black box for us. Furthermore, we will need to ask whether we do not in fact give the uncontrolled power to the authors of the used algorithms by involving artificial intelligence in the above-mentioned decision-making.
In the paper, we focus on the consequences of the involvement of artificial intelligence in deciding political questions for democratic legitimacy and the concept of constitutionalism. We will also analyse the limits that should be set by future legal regulations in this context.