Italian Rules on AI as a supplement to the AI Act

On 20 March 2025, Italian Senate approved bill 1146/2024 ('AI Bill'), which delegates the Government to adopt, within 12 months, a legislative decree to adapt national legislation to Regulation (EU) 2024/1689 ('AI Act'). The AI Bill therefore passes to the Chamber of Deputies for examination.  The AI Bill, recently approved by the Senate, is not intended to overlap with the AI Act - most of whose definitions it recalls - but to accompany its regulatory framework.

General principles

The AI Bill begins by highlighting general principles rooted in respect for fundamental rights and freedoms under the Italian Constitution and EU law. These principles include transparency, proportionality, security, personal data protection, confidentiality, accuracy, non-discrimination, gender equality, and sustainability.

The AI Bill establishes key obligations for the development of AI systems and models, emphasising the importance of respecting human autonomy and decision-making. It asserts that AI must not undermine the democratic structure of institutional and political life.

The AI Bill appropriately underscores the significance of cybersecurity to safeguard the proper functioning of AI systems and emphasises the principle of accessibility.

It reaffirms the protection of freedom of information, personal data confidentiality, and the privacy of minors. Furthermore, it outlines goals for promoting and advancing AI technologies.

For public administrations, the Bill prioritises AI solutions provided by third-party vendors who ensure the localisation of data centres within national territory. These solutions must also meet high standards of security and transparency. In matters of national security and defense, similar rules are established. Specifically, AI systems intended for use in the public domain—excluding those deployed abroad for military operations—must be hosted on servers located within national territory.

Sector-specific provisions

The AI Bill includes detailed provisions tailored to specific sectors, as outlined below:

Health and disability

The use of AI systems in healthcare must not condition access to services based on discriminatory criteria.

Individuals must be informed about the use of AI systems that concern them.

Decisions related to prevention, diagnosis, treatment, and therapeutic choices must remain under the authority of medical professionals.

AI systems and the data they use must be reliable, regularly monitored, verified, and updated.
Specific regulations address electronic health records, health surveillance systems, and the governance of digital health. Among the measures is the creation of an AI platform led by AGENAS, the national agency for digital health.

Research and experimentation

Rules are established for data processing in research and experimentation involving AI and machine learning. This includes scientific experimentation for the development of AI systems in healthcare, targeting prevention, diagnosis, treatment, drug development, therapeutic technologies, rehabilitation tools, and medical devices (including prosthetics and interfaces that support patients). The Minister of Health is tasked with issuing a decree within four months regarding personal data processing for research and experimentation purposes.

Labour Law

The AI Bill affirms general principles protecting workers. Employers are required to inform employees about the use of AI systems. Additionally, an Observatory on the Adoption of AI Systems in the Workplace will be established under the Ministry of Labour and Social Policies to define strategies for integrating AI into the work environment.

Intellectual professions

The use of AI systems is permitted solely for tasks that support and complement professional activities (however, the intellectual work itself must remain the predominant aspect of the service provided). Clients shall be informed about the AI systems used by the professionals.

Public administration and judicial activity

Specific rules are outlined regarding the use of AI within public administration and judicial processes.

Additional provisions in the AI Bill

The AI Bill specifies rules for the application of AI within public administration and judicial processes. Amendments are also proposed to the code of civil and criminal procedure and to copyright law.

The Bill ensures copyright protection for works of creative human genius, even if created with the assistance of AI tools, provided they result from the author’s intellectual effort. Moreover, Article 70-septies is introduced into copyright law (Law 633/1941). Under this article, reproductions and extractions of works or other materials contained within legitimately accessible networks or databases are permitted for the extraction of text and data using AI models and systems, including generative ones, as long as they comply with Articles 70-ter and 70-quater of the same law.

Within 12 months, the Italian Government is tasked with adopting a legislative decree to establish uniform regulations concerning the use of data, algorithms, and mathematical methods for training AI systems.

The AI Bill designates the Agency for Digital Italy (AgID) and the Agency for National Cybersecurity (ACN) as the principal national authorities overseeing artificial intelligence. This is without prejudice to the roles of the Bank of Italy, CONSOB, and IVASS as market supervisory authorities.

Finally, we await the Italian Government’s publication of the legislative decree aimed at aligning national legislation with the AI Act and addressing the further measures outlined in the AI Bill.

Our team is available to provide you with updates on the implementation measures of the AI Act at the national level as well as assistance regarding these measures.