Law 132 of September 23, 2025, on artificial intelligence has been published in the Official Gazette, with Chapter VI dedicated to the relationship between AI and copyright.
In particular, the Italian law addressed two aspects that had previously been the subject of reflection and debate: the protectability of AI-generated creations under copyright law and the possibility of using protected intellectual works for the purpose of training AI systems.
Article 25, paragraph 1, letter a) of the Law provides a negative answer to the first question, establishing that copyright protection is reserved for works of human intellectual creation, specifying, however, that the use of AI in the creative process is permitted, but only as an auxiliary function to the author's intellectual work.
As for the second aspect, Article 25, paragraph 1, letter b) establishes that the extraction of text and data from protected works (Text and Data Mining, according to the wording of the digital copyright directive) using AI systems is permitted, albeit within the strict limits set out in Articles 71-ter and 71-quater of the Italian Copyright Law (introduced precisely in transposition of EU Directive 2019/790).
By virtue of this provision, text and data mining (TDM) may be carried out by research organizations and cultural heritage institutions for scientific research purposes, using works or other materials available in networks or databases to which they have lawful access.
Outside of this scenario, TDM will be permitted provided that the rights holder has not exercised the so-called “opt-out,” i.e., has not expressly reserved the use of the protected works and materials.
Beyond the wording chosen by the national legislator, it will be necessary to understand, on the one hand, what is meant by AI auxiliary support for the creation of the work and, on the other, whether the extension of the TDM exception to AI training activities tout court will be considered in line with the (mandatory) European rules on exceptions and limitations to copyright.
The discussion therefore seems destined to remain essentially open on both questions.
