Artificial Intelligence in Italian Public Administration. Challenges, case studies and regulatory perspectives^**This is a revised version of the paper presented during the Workshop “Künstliche Intelligenz in der öffentlichen Verwaltung Probleme, Praxisbeispiele, Regelungsvorschläge" held at the University of Vienna, Department of Law, on 22 April 2022. The author is grateful to Professors Iris Eisenberger and Franz Merli for organizing this very stimulating workshop and for the kind hospitality. She is also thankful to the many participants for their valuable insights.

In Italy, the employment of AI for decision-making in the context of administrative procedures is increasingly common, being based on the claimed gains in efficiency and neutrality. Local and central public administrations use AI to streamline not only their services to citizens, but also their supervisory activity, eg, in the detection of money laundering or in their controls against tax evasion. However, to date, the use of AI by Public Administrations in Italy is only recognised at the level of policy strategies or, at most, in highly delimited regulatory interventions, while a specific regulatory framework is still missing. Moreover, although robotic decision-making has repeatedly been the subject of administrative case law, conflicting interpretations have taken place over the past few years.