Italy's Consiglio di Stato on Algorithmic Teacher Assignment — Knowability, Human Oversight and Non-Discrimination as Binding Rules

Following Law 107/2015, Italy's education ministry used an algorithm to assign teaching posts in the extraordinary hiring plan and the 2016 mobility procedure. Teachers who found themselves behind colleagues with worse rankings asked to see how the algorithm worked, and challenged the outcomes in the administrative courts.
In judgments of 8 April 2019 (no. 2270) and 13 December 2019 (no. 8472), the Consiglio di Stato accepted that algorithms can be used in administrative procedures, but only with safeguards. It set out three principles: knowability, meaning people have a right to know that an automated process exists and to receive meaningful information about its logic; non-exclusivity, meaning a decision with legal effects cannot rest solely on automated processing and human oversight must be able to validate or reject it; and non-discrimination, with the algorithm's logical correctness and lawfulness open to judicial review. The court drew on GDPR Article 22 and Recital 71.
The ministry's appeals were rejected because the algorithm's opacity breached these principles. The rulings are regarded as landmark decisions on administrative algorithms. Their limit is that they work through individual litigation: they set a standard but do not themselves create a public register or audit duty.

Read the full analysis: https://www.eius.it/giurisprudenza/2019/760