ItalyCodice Penale Art. 437-bis, Omitted Safety Measures in High-Risk AI Systems and Unlawful Alteration of Them
Codice Penale, art. 437-bis, inserted by D.Lgs. 9 settembre 2026, n. 160, art. 12
In force since .
An AI risk obligations rule binding public and private bodies.
- Obligation class
- Governance, Security
- Audit expectation
- none
As of .
What it requires
- Adopt the technical safety measures provided for the design, training, production and placing on the market of high-risk AI systems, suited to prevent malfunctions or alterations of how the systems work, and adopt human-oversight measures, where omitting them would endanger life or public or individual safety, or the security of the State (Codice Penale art. 437-bis, first paragraph).
- If you are a professional user of a high-risk AI system, do not deliberately omit human-oversight measures where the omission would endanger life or public or individual safety, or the security of the State (Codice Penale art. 437-bis, last paragraph).
- Do not alter a high-risk AI system where danger to life or to public or individual safety, or to the security of the State, results (Codice Penale art. 437-bis, second paragraph).
If you get it wrong
Criminal exposureYes
Criminal exposure note
Imprisonment from one to five years for omitting the technical safety measures or the human-oversight measures for high-risk AI systems where danger to life or to public or individual safety results, and from two to eight years where danger to the security of the State results. Imprisonment from two to six years for altering a high-risk AI system where danger to life or to public or individual safety results, and from three to ten years where danger to the security of the State results. The penalty for an omission committed through gross negligence is reduced from one third to one sixth. A professional user who deliberately omits human-oversight measures is punished with the penalties for the omission.
What it makes you log
Logging duty
Article 437-bis punishes omitting the technical safety and human-oversight measures for high-risk AI systems, and altering such systems. It states no duty to create, keep or produce a log or record of any kind.
- Kind
- None
- As of
What this law does
Article 12 of Decreto Legislativo 160/2026 inserts Article 437-bis into the Codice Penale (Penal Code), headed omitted adoption of safety measures in AI systems and unlawful alteration of the systems.
Whoever omits the technical safety measures provided for the design, training, production and placing on the market of high-risk AI systems, suited to prevent malfunctions or alterations of how the systems work, or omits human-oversight measures, is punished with imprisonment from one to five years where danger to life or to public or individual safety results. Where danger to the security of the State results from the omission, the penalty is imprisonment from two to eight years.
Unless the act is a more serious offence, whoever alters a high-risk AI system outside those cases is punished with imprisonment from two to six years where danger to life or to public or individual safety results. Where danger to the security of the State results from the alteration, the penalty is imprisonment from three to ten years. If a first-paragraph omission is committed through gross negligence, the penalty is reduced from one third to one sixth.
A professional user of a high-risk AI system who deliberately omits human-oversight measures is punished with the penalties of the first paragraph, first and second sentences, according to whether danger to life or to public or individual safety, or danger to the security of the State, results.
When LexLint raises it
When your app profile says your app makes high-risk automated decisions, trains models or distributes a software product.