ARTIFICIAL INTELLIGENCE AND CRIMINAL LAW: the news introduced by the Italian law 132/2025

Tradotto da Elena Ciullo

Approved on 23rd September 2025 and come into force as of 10 October, the Italian Law No. 132 marks a turning point: it is the first Italian legislation to comprehensively regulate the use of artificial intelligence. It is part of the European framework outlined by the AI Act, but with a historic particularity: the introduction of new offences and the possibility of using AI systems in criminal proceedings.

The law establishes principles for research, experimentation and application of artificial intelligence systems, promoting their correct, transparent and responsible use. The underlying idea is clear: artificial intelligence can be a useful tool for humans, but it must never replace them.

This also applies in trials. Article 15 entrusts the judge with all decisions regarding the interpretation of the law, the assessment of facts and the adoption of measures, even in cases where AI systems are used. The Ministry of Justice, on the other hand, will be responsible for regulating the use of artificial intelligence to improve the organisation of services and simplify judicial work.

In addition, the law delegates the Government to adopt, within twelve months, legislative decrees to define the rules on the use of data, algorithms and mathematical methods for training AI systems. The aim is to create a comprehensive framework, identifying rights, obligations and penalties for those who use such tools.

In criminal law, the law directly affects the code: aggravating circumstances are introduced for crimes committed "through, with or by means of" artificial intelligence systems, when these make the conduct more insidious or hinder the defence.

Article 612-quater of the Criminal Code is then added, which punishes with imprisonment from one to five years anyone who disseminates, without consent, images, videos or voices falsified with AI systems, capable of misleading others as to their authenticity. This is the first offence in the Italian Criminal Code dedicated to deepfakes: realistic but false content, created using generative artificial intelligence and without the consent of the subjects represented, which reworks and adapts images and videos to contexts other than their original ones.

This step follows the example set by Denmark, the first country in Europe to take such action. At a time when deepfakes can also be used as political tools, capable of spreading misinformation or altering public debate, this is a necessary safeguard.

The legislator also intervenes in the field of copyright, extending protection to works reproduced or processed by artificial intelligence systems without the author's consent.

However, one crucial question remains: if an AI system causes damage, who is liable? The programmer, the user or the entity that employs it? For now, the law does not offer precise answers and leaves it to delegated decrees to define the criteria for attributing liability.

With Law 132/2025, Italy officially enters a new era of justice, not only criminal justice. This is a necessary law that tries to keep pace with reality and regulate a constantly evolving phenomenon. But it is also a starting point: it is essential to open a public debate, not only among lawyers and experts, but among all citizens, on how much technology can and should enter the most sensitive area of law, that of criminal proceedings, which can limit the most important thing we possess: freedom.

Mondo Internazionale APS – Riproduzione Riservata © 2025