The Italian Council of Ministers has approved the draft law on artificial intelligence, laying down principles on research, experimentation, development, adoption and application of AI systems and models (“Draft Law”).

As of today, the Draft Law has been adopted by the Italian Government and presented to the Italian Parliament for discussion.

Below we provide a brief summary of the 26 articles that comprise the Draft Law.

In a nutshell, the Draft Law is aimed at regulating the proper, transparent, responsible and anthropocentric use of AI in Italy within the European and Italian regulatory framework. This is one of the very first examples of primary legislation to be read in conjunction with the AI Act. If you want to find out more about the Italian landscape, check out our article on the Italian strategy on AI 2024–2026.

Principles

The Draft Law emphasizes that the research, experimentation, development, adoption, application and use of AI systems and models must respect the fundamental rights and freedoms enshrined in the Italian Constitution and EU law, as well as the principles of transparency, proportionality, security, data protection, confidentiality, robustness, accuracy, non-discrimination, gender equality and sustainability, and that the use of AI should improve the nation’s economic competitiveness.

Among the other things, the Draft Law clarifies the following:

• Cybersecurity should be ensured throughout the entire lifecycle of AI systems and models. Specific security controls should be adopted to ensure their resilience against attempts to alter their use, expected behavior, performance or security setting.
• The development of AI systems and models must be grounded on data and processes that are correct, reliable, secure, high-quality, appropriate and transparent. Furthermore, the use of AI must uphold the freedom and pluralism of the media, the freedom of expression, and the objectivity, completeness, impartiality and fairness of information. It goes without saying that the GDPR (Reg. (EU) 2016/679) and Italian data protection laws should be complied with at all times.
• Access to AI by minors under 14 years of age requires the consent of the person exercising parental responsibility. Minors over 14 years can express their consent for the processing of personal data related to AI so long as the information and communications are easily accessible and understandable.

Sector-specific provisions

The Draft Law upholds a sector-specific approach, declining rules applicable to the use of AI in various fields, such as health, work, justice and public administration. A few examples are below:

• Healthcare sector – Patients must be informed about the use of AI and its advantages, and doctors should always have the final say on medical treatments and therapy options, without relying solely on AI outputs.
• Employment sector – Any use of AI that may result in discrimination of the worker is banned. All workers must always be informed about the use of AI.
• Intellectual professions – AI must only be used for collateral and non-core activities. Professionals (e.g., lawyers, architects, etc.) must inform the recipient of their services about the use of AI systems and their functioning, using clear, simple and complete language.

The Authorities

According to the Draft Law, the Agency for Digital Italy (Agenzia per l'Italia Digitale, AgID) and the National Cybersecurity Agency (Agenzia per la cybersicurezza nazionale, ACN) are to be appointed as the national authorities for AI, and they will be entrusted with adequate powers and responsibilities to ensure the application and implementation of national and European legislation on AI. That said, this appointment has currently generated debate among public institutions as some say that this role should be entrusted to independent authorities. Click here to check out our article on the current position of the Italian Data Protection Authority (“Garante per la protezione dei dati personali”).

Copyright

The Draft Law suggests amending the Italian Copyright Law (Law n. 633/1941), by adding copyright protection for works created by AI systems, as long as human input was inventive, significant and can be proven. This does not affect the rights of the developer of the AI system used to generate protected works.

• Furthermore, the Draft Law provides that the reproduction and extraction of works or other materials through AI are allowed, in accordance with the existing provisions for text and data mining (e.g., provided that the use of works or materials whose use is expressly reserved by the rights holders is avoided).

Italian Civil Code and Italian Criminal Code

• The Draft Law amends the Italian Civil Code and the Italian Criminal Code. With specific reference to the latter, the Draft Law introduces the use of AI systems as a common aggravating circumstance, increases some sanctions (e.g., for the crimes of substitution of person, fraud, and money laundering) and introduces the new crime of “unlawful dissemination of AI-generated or manipulated content.” 

Next steps for companies

We are still at a very early stage and we expect the Draft Law to be amended and further refined during the parliamentary discussion to avoid unnecessary overlapping with the AI Act.That said, it will be important for companies to monitor the Draft Law as it may set an example providing meaningful guidance for navigating the regulatory landscape surrounding AI deployment in other areas of law (e.g., criminal law, copyrights, etc.) and/or other jurisdictions.