In 2025, Italy became the first EU Member State to adopt a comprehensive national framework on artificial intelligence, following the adoption of the European Union’s AI Act. Law No. 1146-B, passed in the spring of 2025, sets out a national strategy for AI, strengthens institutional governance, regulates sensitive sectoral uses, and introduces new criminal offences. The law reflects a clear ambition: to reconcile technological innovation with legal certainty, while ensuring consistency with European Union law.
Law No. 1146-B was approved by the Italian Senate on 20 March 2025 and subsequently adopted by the Chamber of Deputies on 25 June 2025. It will enter into force after promulgation and publication in the Italian Official Gazette. The law provides for a twelve-month period for the adoption of implementing legislative decrees to define technical and organisational details.
The Italian law adopts an anthropocentric approach to artificial intelligence. Its aim is to promote research and innovation while ensuring that development and use remain transparent, accountable, and respectful of fundamental rights. AI is viewed as a tool for efficiency and modernisation, provided that ultimate responsibility always rests with human decision-makers. The law also highlights the importance of anticipating the economic, social, and ethical risks associated with the deployment of AI systems.
The national AI strategy is coordinated by the Presidency of the Council of Ministers, subject to review by the Interministerial Committee for Digital Transition. Two agencies are formally designated as national competent authorities: the Agenzia per l’Italia digitale (AgID), responsible for innovation and notification functions, and the Agenzia per la cybersicurezza nazionale (ACN), in charge of supervision, inspections, and sanctions, particularly with regard to cybersecurity aspects. These agencies also act as the designated supervisory authorities under the EU AI Act, in cooperation with the Bank of Italy, CONSOB, and IVASS to ensure effective sectoral oversight in banking, financial, and insurance markets.
The law regulates the use of AI in several sensitive sectors. In the field of labour, it prohibits discriminatory practices and imposes an obligation to inform workers. Within public administration, it promotes the use of AI to increase efficiency and transparency, while confirming that decision-making remains the sole responsibility of human agents. In the justice sector, it authorises controlled experimentation and mandates digital training for judges and court staff. Specific provisions also govern the use of AI by law enforcement and security services, ensuring compliance with fundamental rights.
The repressive dimension of the law is particularly innovative. It introduces a new criminal offence for the unlawful dissemination of deepfake content, increases penalties for offences committed through the use of AI, and establishes civil liability rules that include heightened safety obligations for operators and a partial reversal of the burden of proof for certain high-risk systems. Copyright law is also affected, with provisions regulating text and data mining by AI, in line with Articles 70-ter and 70-quater of Italian law.
A central aspect of Law No. 1146-B is its explicit alignment with the European Union’s AI Act (Regulation (EU) 2024/1689). The Italian provisions must be interpreted consistently with the EU framework. The designated national authorities, AgID and ACN, are entrusted with the functions assigned to Member States under the AI Act. Furthermore, the government is empowered to adopt, within twelve months, legislative decrees aimed at harmonising national law with the EU regulation, particularly concerning system notification, conformity assessments, and sanctions. Far from competing with the European framework, the Italian law is conceived as an instrument of support and anticipation, designed to facilitate the effective implementation of the AI Act within Italy.
