Home

Internal

Global

Contact

de|en|it

Rödl & Partner Italy: 2024 turnover grows to a record level of 33.3 million euros |

Worldwide

We use cookies to personalise the website and offer you the greatest added value. They are, among other purposes, used to analyse visitor usage in order to improve the website for you. By using this website, you agree to their use. Further information can be found in our data privacy statement.

Italy leads the way: the first national law on Artificial Intelligence is approved

published on 29 September 2025 | reading time approx. 8 minutes

With its publication in the Official Gazette on 25 September 2025, Law No. 132 of 2025 (hereinafter, the “Italian AI Act”) officially becomes part of the Italian legal system, with entry into force scheduled for 10 October 2025. The statute, representing the first comprehensive Italian legislation on artificial intelligence, is the outcome of a complex and articulated parliamentary process: from Bill No. S.1146 presented in the Senate, to the subsequent approval with amendments of Bill C.2316 by the Chamber of Deputies, followed by the Senate’s final adoption of Bill S.1146-b, which became law upon its publication in the Official Gazette.

This legislative intervention comes at a crucial moment for the development of artificial intelligence (AI) technologies, in the wake of the adoption of Regulation (EU) 2024/1689 (the so-called “European Union AI Act”), which is set to become the European—if not global—regulatory benchmark. Through this statute, Italy seeks not only to harmonise domestic law with EU legislation but also to reinforce certain fundamental principles, ensuring that AI is deployed in accordance with the criteria of transparency, proportionality, safety, the protection of fundamental rights, and the centrality of the individual.

Law No. 132/2025 is structured along two main axes: on the one hand, it lays down general principles applicable to all AI systems and models; on the other, it introduces specific provisions for key sectors such as healthcare, research, employment, the liberal professions, public administration, justice, copyright, and criminal law. The statute is further supplemented by a series of legislative delegations empowering the Government to issue implementing decrees and regulations.

The purpose of this analysis is to illustrate the main contents of the Act, highlighting its practical implications and the critical issues emerging from scholarly and professional debate, and concluding with operational recommendations addressed to businesses, professionals, and public administrations.

The Italian AI Act “in a nutshell”

General Principles and Fundamental Rights

Title I of the Act establishes that research, development, and application of AI must comply with constitutional rights, EU law, and the principles of transparency, proportionality, security, personal data protection, non-discrimination, gender equality, and sustainability (Art. 3). Central to this framework is the guarantee of human oversight: no decision may be entirely delegated to an automated system. The democratic process is expressly safeguarded, prohibiting any use of AI that may compromise political debate or the sovereignty of the State.

With regard to data and information (Art. 4), the Act reaffirms the primacy of GDPR principles: lawfulness, fairness, transparency, and purpose limitation. Special attention is devoted to minors: under the age of 14, the use of AI systems requires parental consent; between 14 and 18 years of age, minors may consent autonomously, provided that information notices are comprehensible.

Specific areas of application

Healthcare and Scientific Research (Arts. 7–10): AI is recognised as a tool to support diagnosis and treatment, but the final decision remains with the physician. Discrimination in access to healthcare services is strictly prohibited. The Act enhances the role of AGENAS, entrusted with managing a national AI platform for territorial healthcare services and with issuing guidelines on data anonymisation and synthesis, subject to the opinion of the Data Protection Authority. Secondary use of personal data, including special categories, is authorised for scientific research purposes, provided that such data are anonymised or pseudonymised.
Employment (Arts. 11–12): AI may be used in employment relationships, subject to the protection of workers’ rights. Employers must provide clear information on the logic and purposes of AI systems and ensure human intervention in automated decision-making. A National Observatory on AI is established within the Ministry of Labour to monitor the impact of AI on employment and workers’ rights, with consultative and advisory functions.
Liberal Professions (Art. 13): AI may support professional activities but cannot replace the creative and intellectual contribution of the professional. Professionals must also inform clients of any use of AI systems, in order to safeguard transparency and the fiduciary relationship.
Public Administration (Art. 14): AI may be employed to enhance efficiency and quality of public services, subject to obligations of traceability and intelligibility of system operations. Personal liability of public officials remains unaffected.
Justice (Art. 15): AI may be used in judicial proceedings solely for organisational purposes and case-law research. Delegation of judicial decision-making to algorithms is strictly prohibited: fact-finding and legal interpretation remain the exclusive competence of the judge.
Copyright (Art. 25): Works created with the assistance of AI may be protected by copyright only if they result from a significant human creative contribution. Content generated or manipulated by AI must be labelled to ensure traceability.
Criminal Law (Art. 26): Use of AI constitutes an aggravating factor for criminal offences. A new offence is introduced: the unauthorised dissemination of misleading deepfakes (Art. 612-quater of the Criminal Code), punishable by imprisonment from 1 to 5 years. Aggravating factors are also provided for fraud, corporate crimes, and market abuse committed through AI.

Government Delegations

Article 24 of Law No. 132/2025 grants the Government a broad delegation, requiring the adoption, within twelve months of entry into force, of one or more legislative decrees to align national legislation with Regulation (EU) 2024/1689 (the EU AI Act) and to regulate lawful and unlawful uses of AI systems.

The delegation encompasses, inter alia:

General normative alignment: amendment, integration, or repeal of national legislation (including in banking, finance, insurance, and payments) to ensure full transposition of the EU Regulation, granting to national authorities (AgID, ACN, and sectoral regulators) all supervisory, inspection, and sanctioning powers provided by the EU AI Act;
Secondary legislation and sanctions: regulatory powers for independent authorities in specific areas; attribution of sanctioning powers in line with Article 99 of the EU AI Act;
Training and awareness: mandatory training programmes on AI for citizens, professionals, students, and operators, including through professional associations and chambers. Provision may be made for fair remuneration in relation to risks and responsibilities linked to AI use;
Specific sectors: tailored regulation for AI use in law enforcement activities;
Market supervision: power to request information, conduct inspections (including without prior notice), and audit high-risk AI systems;
Sanctioning framework: update of national administrative and criminal sanctions in line with the EU AI Act.

Implementation Timeline

15 days from publication: the Act enters into force on 10 October 2025;
90 days: decree of the Ministry of Labour establishing the AI Observatory;
120 days: decree of the Ministry of Health on the use of data for research and experimentation;
12 months: implementing legislative decrees on data, algorithms, liability, and sanctions;
Transitional phase: the Ministry of Justice may authorise AI experimentation in courts.

Conclusions and Operational Recommendations

Law No. 132/2025 marks a significant milestone: for the first time, the Italian legislator (ahead of other Member States) has established an autonomous legal framework for AI, in coordination with European law. Nevertheless, interpretative uncertainties remain:

the broadening of obligations beyond the EU AI Act (which, being a Regulation, is directly applicable);
the lack of distinction among AI stakeholders (providers, users, distributors) and between categories of AI risk (unlike the EU AI Act, the Italian statute does not appear to differentiate, e.g., between high-risk and minimal-risk systems);
the vagueness of certain provisions, such as information duties in professional contexts and the interplay with copyright and privacy rules.

Accordingly, in order to avoid violations and related sanctions, businesses, public bodies, and professionals should promptly:

Conduct internal mapping: identify AI processes and systems already in use (HR, research, healthcare, PA, citizen services) and classify them according to the risk criteria set out in the EU AI Act;
Implement organisational adjustments: draft clear information notices, adopt internal procedures for human oversight, establish traceability and audit mechanisms, and update compliance programmes (including Model 231) to cover risks arising from unlawful AI use;
Promote training and corporate culture: foster AI-related legal awareness and compliance training, engaging also senior management;
Monitor regulatory developments: closely follow implementing decrees in the coming 12 months to update contracts, processes, and corporate policies accordingly;
Strengthen data governance: enhance data protection procedures, particularly anonymisation and pseudonymisation, in line with AGENAS guidelines and the Data Protection Authority’s requirements.

Ultimately, Law No. 132/2025 serves as a testing ground for the Italian legal and economic system. Its effectiveness will depend not only on the quality of forthcoming implementing decrees but also on the ability of businesses, professionals, and public administrations to interpret and apply it in a coherent, responsible, and innovative manner.