Legal scholars, judges, regional policymakers and AI experts spent a day in Bologna working through a question that keeps surfacing across European administrations: what does it actually mean to govern with AI?

The morning built the technical and legal groundwork, opening with the distinction between deterministic and non-deterministic systems and what that difference means once an administrative decision is at stake. Riccardo Rovatti covered the underlying technology, Giovanni Sartor examined bias and algorithmic opacity as legal problems, and Enrico Carloni set out the principles at issue when administration becomes artificial. Judge Anna Corrado of TAR Lombardia then turned to procurement, reading AI acquisition against Article 30 of Italy’s new public contracts code.

The afternoon moved to practice. AI4Gov-X Scientific Director Gianluca Misuraca presented the project and the MagiX Master’s programme, setting out how public-sector professionals across Europe are being prepared for this work. A team from Regione Emilia-Romagna followed with their own path from experimentation to operational value, and Gianluca Gardini closed the programme on where human judgement has to remain in public decision-making.

Human oversight, responsibility for algorithmic decisions and the skills gap behind both ran through the whole day. That is the ground AI4Gov-X was built to cover.