The training day — with Avv. Alessandro del Ninno as sole speaker — provided a systematic overview of the new European Regulation on artificial intelligence, starting from its entry into force, the relevant application timelines, and the subject matter and objectives of the Regulation, before moving on to examine the main categories of AI, training methods, the distinction between AI models and AI systems, the subjective scope of application and the legal effects of the new rules.

Particular attention was devoted to the relevant legal, technological, commercial and ethical definitions, including those concerning the design, use and operation of AI systems, the actors involved in the value chain, commercial activities relating to AI systems, and systems using biometric technologies.

The programme addressed the core of the Regulation, namely the risk-based approach adopted by the EU legislator. Prohibited artificial intelligence practices were examined, also in light of the European Commission’s implementing guidelines, together with the classification of high-risk AI systems pursuant to Article 6 and Annex III of the Regulation. Avv. Alessandro del Ninno guided participants through the analysis of the compliance requirements applicable to high-risk AI systems as a prerequisite for their placing on the market or putting into service, with specific reference to the structuring, implementation, documentation and maintenance of a risk assessment system.

Significant attention was given to the practical obligations relating to risk assessment, data governance, technical documentation, logging and retention of logs, transparency towards users, cybersecurity, accuracy, robustness, resilience and human oversight. The horizontal obligations incumbent upon providers of high-risk AI systems were also examined, including the implementation of a quality management system as a form of artificial intelligence accountability, as well as conformity assessment procedures and the obligations imposed on the other actors involved in the value chain, with a specific focus on deployers.

The second part of the conference was devoted to the managerial, informational and contractual aspects connected with the use of AI systems, the impact on supply contracts, the drafting and documentation of the Fundamental Rights Impact Assessment for high-risk AI systems, as well as the obligations relating to general-purpose AI models, including GPAI models and models with systemic risk.

The programme also explored copyright protection, the regulation of deepfakes, the new rights of natural persons in the context of interaction with AI systems, and regulatory sandboxes as legal instruments for the early-stage testing of technological AI solutions.

The day concluded with an analysis of European AI governance, the role of the AI Office, the European Artificial Intelligence Board, experts, the AI Forum and national authorities, as well as an examination of codes of conduct, the sanctions regime, liability and compensation for damage caused by AI systems, and initial references to Italian AI legislation.