The Law sets national principles and sector-specific rules for AI research, development, provision and use, designates national AI authorities, preserves human responsibility in sensitive decisions, and introduces enforcement and criminal-law provisions.
Regulation
Verified record
- Risk classification summary
- The Law does not create a general horizontal AI risk-classification system; it applies cross-cutting principles and sector-specific safeguards.
- Requirements summary
- AI research, development, provision and use must respect transparency, proportionality, security, data protection, non-discrimination and human oversight. Patients and professional clients must be informed of AI use, medical and judicial decisions remain human, and health-sector AI systems require periodic verification.
- Affected entities
- Public and private actors researching, developing, providing or using AI or general-purpose AI systems in Italy, including health professionals, professional service providers and judicial bodies within the relevant provisions.
- Enforcement summary
- The National Cybersecurity Agency, sectoral authorities and the Italian Data Protection Authority exercise their existing competences under the Law.
- Penalty summary
- Article 612-quater of the Criminal Code provides imprisonment from one to five years for harmful non-consensual AI-generated or altered falsification of a person’s image, voice or appearance.