“Like any human creation, artificial intelligence has its own unknown shadows—bearing seeds of every kind, yet also the sparks of new life.”

While this phrase may sound like the opening line of a new poem, it was actually the introduction to the illustrative report attached to the new Italian AI bill (no. 1146) which was approved, on March 20, 2025, by the Italian Senate.

The bill, which is now moving to the Italian Chamber of Deputies for final approval, aims to integrate within the national legal framework provisions that go beyond those already implemented at the European level with the AI Act.

In particular, the new legislation aims to address specific areas such as data governance, national security, healthcare, labor, and intellectual property, in order to further promote the development of new technologies. The Government has allocated a €1 billion investment fund to support AI and related technologies, underlining its ambition to remain competitive in the rapidly evolving tech landscape.

Let’s take a closer look at the key provisions of the bill to (try to) understand which “seeds and sparks” AI may bring according to the Italian Government.

***

At its core, the bill establishes a governance framework led primarily by the Agency for Digital Italy (AgID) and the Italian Cybersecurity Agency (ACN). The proposed framework has been criticized by some commentators, who see the Italian Data Protection Authority as the most suitable (and independent) candidate for overseeing AI rules. It seems now that the AgID and ACN will likely be confirmed as the authorities in charge for steering the AI rudder in Italy.

The bill provides general legal provisions and principles, delegating to the Italian Government and other bodies/institutions the responsibility for drafting a more comprehensive framework. For example, the use of AI by professionals should always be ancillary and disclosed to clients, but the Italian Government will set out specific rules for the regulation of data usage in AI training at a later stage through ad-hoc legal acts.

That said, the government has already taken a firm stance on some specific areas.

As for national defense, the bill lays down the foundation for a new cybersecurity and defensive strategy. Importantly, AI intended for use in the public sector shall operate on servers located within Italy to strengthen national security and protect the safety of citizens’ data.

In the healthcare sector, AI is expressly recognized as a tool to support - but not replace - medical professionals. AI systems can indeed operate as platforms to assist healthcare providers with non-binding recommendations to improve diagnosis and treatment. The bill also sets up the first steps for the development of a public AI system in support of healthcare professionals. Patients will have the right be informed about the use of such systems.

Furthermore, the bill recognizes that processing health data for research purposes by non-profit public and private entities falls within the exemption provided by art. 9.2.g of the GDPR, exempting de facto data controllers from collecting individual consent. Secondary use is largely allowed as long as specific conditions apply and privacy notices are made available to data subjects (also through documents published on a website). A specific communication should be made to the Italian Data Protection Authority when processing health data for research purposes. 

Moving on to workplaces, AI is meant to enhance working conditions and productivity while ensuring transparency and protecting workers’ dignity. Employers are required to inform employees about the usage of AI in automated decision-making or monitoring systems.

The bill also regulates AI-generated content. In particular, works created with the aid of artificial intelligence tools can also be protected by copyright, provided that their creation “results from the intellectual effort of the author”.

In the public administration and judiciary field, the bill recognizes that AI can be used to streamline processes, but public representatives and judges shall retain full decision-making authority.

***

Whereas from a practical standpoint, the bill provides some new important rules that should be taken into consideration by entities operating in Italy, this new piece of legislation fits within the Italian Government’s larger discussion around AI. It is still hard to say with certainty what “seeds” the increased use of AI will bring, and whether the current bill will be further amended. In any event, Italy is stimulating a debate to ensure that these “seeds” grow into sustainable solutions that benefit society without compromising fundamental rights or ethical standards. 

At the end of the day, isn’t this the critical issue for which that all Governments (not just Italy’s) and operators will be held accountable in the coming years?