Newsletter November 2026
Artificial Intelligence and Corporate Liability
Changes introduced by Legislative Decrees No. 160/2026 and No. 179/2026: new predicate offences, AI system safety and implications for compliance under Legislative Decree No. 231/2001
1. Introduction
The framework governing the administrative liability of entities is increasingly required to address the new risks arising from the development and use of artificial intelligence (AI) systems.
Pursuant to the legislative powers delegated by Law No. 132 of 23 September 2025, Legislative Decree No. 160/2026 (Official Gazette No. 214 of 15 September 2026) (“Decree”), concerning the use of AI in policing and civil and criminal liability, and Legislative Decree No. 179/2026 (Official Gazette No. 234 of 8 October 2026), concerning the powers of national authorities and the use of AI in education and training, have been enacted. Both decrees have entered into force and form part of the national legislative framework implementing Regulation (EU) 2024/1689 (“AI Act”).
The two legislative decrees address distinct areas:
- Legislative Decree No. 160/2026, which is particularly relevant to the administrative liability of entities, introduces provisions concerning criminal and civil liability. In particular, Article 12 introduces the new Article 437-bis into the Italian Criminal Code, concerning the failure to adopt safety measures in AI systems and the unlawful alteration of such systems, whilst Article 15 expands the list of predicate offences under Legislative Decree No. 231/2001 by introducing the new Article 25-vicies, concerning offences committed through the use of AI systems;
- Legislative Decree No. 179/2026 further develops the national framework for implementing the AI Act, regulating, amongst other matters, the powers of the competent national authorities. Of particular relevance is Article 42, which introduces specific provisions on occupational health and safety. It requires the use of AI systems affecting work organisation, production rates, working methods or decision-making processes relevant to safety to be assessed as part of the risk assessment provided for under Article 28 of Legislative Decree No. 81/2008. The provision also requires employers to ensure that workers receive adequate information and training on the specific risks associated with the use of such systems and the preventive and protective measures adopted.
2. The new Article 437-bis of the Italian Criminal Code
Article 437-bis of the Italian Criminal Code, entitled “Failure to adopt safety measures in artificial intelligence systems and unlawful alteration of systems”, introduces criminal law safeguards against risks associated with high-risk AI systems. Its placement immediately after Article 437 of the Italian Criminal Code reflects, from a systematic perspective, the preventive approach to safety protection.
The first paragraph provides for imprisonment of between one and five years for anyone who fails to adopt the technical safety measures required for the design, training, production and placing on the market of high-risk AI systems, intended to prevent malfunctions or alterations to their operation. The same penalty applies to the failure to implement adequate human oversight mechanisms, where such omissions give rise to a risk to life or individual or public safety. A more severe penalty applies where the risk concerns State security, in which case imprisonment ranges from two to eight years.
The second paragraph criminalises the unlawful alteration of high-risk AI systems, except where the conduct constitutes a more serious offence and outside the circumstances covered by the first paragraph. Imprisonment ranges from two to six years where the alteration gives rise to a risk to life or individual or public safety, and from three to ten years where the risk concerns State security.
Where the conduct referred to in the first paragraph is committed through gross negligence, the third paragraph provides for a reduction in the applicable penalty of between one-sixth and one-third.
The final version of the provision also introduces a fourth paragraph concerning the liability of professional users of high-risk AI systems. This extends the penalties provided for in the first paragraph to anyone who intentionally fails to implement human oversight measures, where such failure gives rise to a risk to life or individual or public safety, or to State security. The applicable penalties depend on the nature of the risk involved.
3. The new Article 25-vicies and the liability of entities
Article 15 of the Decree introduces the new Article 25-vicies, entitled “Offences committed through the use of artificial intelligence systems”, into Legislative Decree No. 231/2001, thereby expanding the list of predicate offences giving rise to the administrative liability of entities.
With regard to the offence under Article 437-bis of the Italian Criminal Code, the provision establishes a pecuniary sanction ranging from 600 to 1,000 units, subject to the fulfilment of the general conditions for establishing the entity’s liability under Legislative Decree No. 231/2001.
The provision also includes the offence under Article 612-quater of the Italian Criminal Code, concerning the unlawful dissemination of artificially generated or manipulated content, for which entities may be subject to a pecuniary sanction ranging from 200 to 700 units.
The new legislative framework therefore identifies two distinct areas of criminal liability: on the one hand, the safety of high-risk AI systems and, on the other, the use of AI to generate or alter content potentially harmful to individual rights.
The inclusion of these offences amongst the predicate offences under Legislative Decree No. 231/2001 broadens the scope of entities’ liability, allowing the sanctions provided for under the Decree to apply to certain unlawful activities involving artificial intelligence.
However, such liability does not arise merely from the unlawful use of AI technology. Rather, it requires the general conditions for attributing liability under Legislative Decree No. 231/2001 to be met, including the commission of a predicate offence in the interest or to the advantage of the entity by persons holding a relevant position within its organisational structure.
For both offences, Article 25-vicies provides for the disqualifying sanctions set out in Article 9(2)(b), (c), (d) and (e) of Legislative Decree No. 231/2001, namely: the suspension or revocation of authorisations, licences or concessions instrumental to the commission of the offence; a prohibition on contracting with public authorities; exclusion from benefits, financing, grants or subsidies, including the possible revocation of those already granted; and a prohibition on advertising goods or services.
The provision does not refer to Article 9(2)(a), concerning disqualification from carrying out business activities.
4. Implications for compliance under Legislative Decree No. 231/2001
The expansion of the list of predicate offences makes it advisable for entities exposed to the relevant risks to update their risk assessments and review the adequacy of their Organisational, Management and Control Models under Legislative Decree No. 231/2001.
Particular attention should be paid to activities involving the design, training, production, placing on the market and professional use of high-risk AI systems, as well as to processes involving the generation and dissemination of AI-generated content.
From an operational perspective, particular importance should be attached to identifying and mapping risk-sensitive activities, allocating internal responsibilities, implementing controls over technical safety measures and human oversight mechanisms, providing appropriate staff training and ensuring adequate information flows to the Supervisory Body (Organismo di Vigilanza).
These safeguards should be tailored to the entity’s actual activities and the specific risks identified.
In this context, Article 42 of Legislative Decree No. 179/2026 is also particularly relevant, as its provisions require coordination between the management of risks arising from the use of AI and the entity’s occupational health and safety management framework.
Such coordination is especially important when assessing the adequacy of organisational safeguards under Article 25-septies of Legislative Decree No. 231/2001, concerning offences relating to occupational health and safety.
Accordingly, ensuring compliance with Legislative Decree No. 231/2001 requires entities to review their organisational and control measures in light of the new risks associated with the use of artificial intelligence, whilst ensuring consistency with the obligations established under the AI Act and national legislation.
This process should follow a proportionate, risk-based approach, distinguishing between regulatory obligations and the legal conditions for attributing administrative liability to entities.

