This article examines preventive measures from an administrative law perspective as instruments for the anticipatory protection of public security, addressing their constitutional basis, the conditions governing their application, and the framework of safeguards. The analysis also covers anti-Mafia interdiction measures, the shift towards collaborative prevention models, urban security, and the use of new technologies. It highlights the need for a constitutionally oriented model of public security capable of reconciling effective preventive action with the principles of legality, specificity, precision, foreseeability, and proportionality, as well as with effective judicial review.

The article analyses the civil and administrative liabilities associated with the use of Artificial Intelligence, highlighting the humanistic, risk-based approach that characterises European and national legal systems. It examines the main areas of liability and evidentiary challenges related to fault and causality. European regulatory initiatives, including the AI Regulation and legislation on defective products, are also discussed, emphasising the need for systemic interventions capable of ensuring effective protection without hindering innovation. The article also addresses the critical issues related to the potential attribution of legal personality to autonomous artificial intelligence systems. Particular attention is paid to the liability of public administrations in the use of AI systems and the complex nature of administrative liability, which is conditioned both by the technical opacity of such systems and by the regulatory restriction of gross negligence and the fragmentation of the parties involved.