On the legal discrepancy between AI administrative decision-making and fundamental EU administrative principles

Artificial intelligence has started to change the face of public administration by automating the traditional way of exercising administrative functions. This gradual but steady transformation has come with consequences, some of which potentially comprise the established legal and administrative principles and fundamental rights and freedoms protected under EU law. As the role of AI in public decision-making is expanding, it is changing the way administrative discretion is exercised. There is not much administrative law legislation specifically addressing the use of automated decision making (ADM) systems in general and of AI systems in particular, since the dominant role of data protection law in regulating ADM and the idea that any instance of administrative decision-making should comply with certain general principles of administrative law. In the absence of specific legislative provisions administrative courts resort to general principles of administrative law to evaluate the use of ADM by public administrations/authorities. In that respect, the use of AI for making decisions in public administration has prompted serious concerns and sparked an ongoing debate to the point that the legitimacy of AI-involved administrative decision-making has come into question. This article tries to answer this question from the perspective of EU law. After a short review of the academic literature on the subject, the paper analyses the conformity of AI decision-making in public administration with a reference to fundamental values and administrative principles of EU law. The paper seeks to contribute to the scholarly discussion on the legal discrepancy between AI administrative decision-making, fundamental administrative principles, and the existing EU regulation.