Evolving AI-based Automation – The Continuing Relevance of Good Administration

Automated decision-making (ADM) systems using artificial intelligence (AI) are on the rise in EU and Member State administrations. The underlying technology is fast developing. However, requirements for legality and accountability of public decision-making continue to exist. This article argues that ADM systems using AI can be made fit to the requirements of the rule of law and general principles of EU public law. First, their use will require a relatively detailed pre-determination of an ADM system’s capabilities in a legal basis. Second, on the level of decision-making, demanding detailed reasoning especially concerning the information and data used is central for ensuring compliance with EU legal principles. This is a pre-condition for any kind of human review, will contribute to ensuring effective remedies and will be the basis for regular monitoring of the output of ADM.

I sistemi decisionali automatizzati (ADM) che utilizzano l'intelligenza artificiale (IA) sono in aumento nelle amministrazioni dell'UE e degli Stati membri. La tecnologia di base è in rapido sviluppo. Tuttavia, i requisiti per legalità e responsabilità del processo decisionale pubblico continuano ad esistere. Questo articolo sostiene che i sistemi decisionali automatici che utilizzano l'IA possono essere adattati ai requisiti dello stato di diritto e ai principi generali del diritto pubblico UE.