Robotic decisions and administrative discretion: from the black box to the dark matter

Algorithms and big data are used more and more frequently in administrative procedures: to assess civil servants performance, to build credit systems, to allocate resources, to hire people or to profile them as recipients of public programs or to predict criminal proclivity. Robotic decision can be taken through a fully automated process or with a partial human intevention.
Different kinds of robotic decisions, pertaining to different countries and jurisdictions, will be analyzed in the paper,with the main aim to discuss benefits and risks of algorithmic administration.
The focus of the paper will be on three main points:
a. the structure of the automated decision process in different cases;
b. the trasparency and accountability (or lack thereof) of the framework within wich administrative discretion is exercised;
c. the width and depth of judicial review on robotic decisions.