Jump to content

Italy AI Law

From AI Law Wiki

AI-generated text. This page was generated using artificial intelligence.

Italy AI law includes national measures that operate alongside the EU AI Act. Legislative Decree No. 160 of September 9, 2026 sets Italian rules for police use of AI and for civil and criminal matters involving AI systems. It was published in the Gazzetta Ufficiale on September 15 and entered into force on September 30, 2026.[1]

This page summarizes selected provisions of the Italian-language original text, checked October 2, 2026. The decree's application must be read with the EU AI Act and applicable Italian data-protection and criminal-procedure rules; the summaries below do not replace the operative text.[2]

Status and scope

Measure Status and scope
Legislative Decree No. 160/2026 In force from September 30, 2026. Title I addresses police AI systems; Title II addresses criminal provisions and civil procedures for AI-related damages. The decree implements aspects of the EU AI Act under Italy's 2025 AI legislation.[1][2]

Police use and biometrics

Article 1 places police AI use within the EU AI Act and Italy's 2025 AI legislation and states principles of proportionality, nondiscrimination, human oversight and transparency.[3]

Article 8 limits police use of real-time remote biometric identification in public places to specified prevention and search purposes. It requires targeted use, a purpose-specific reference dataset and prosecutorial authorization, with a narrow urgency procedure and deletion requirements if authorization fails.[4] Article 9 requires advance rights and data-protection impact assessments, protected logs and notification to the data-protection authority for these uses.[5]

Article 10 separately regulates retrospective facial recognition using AI components in lawfully installed video-surveillance systems. Targeted searches for a suspected offender generally require authorization from a preliminary-investigations judge, with a stated exception for initial post-crime identification on objective, verifiable grounds. A decision adversely affecting a person may not rest solely on facial-recognition results.[6]

Civil claims involving AI systems

Articles 16–19 address civil damages claims involving AI systems. They provide for court-ordered access to specifically relevant evidence when the claimant makes the claim plausible, subject to proportionality and protection of confidential information. When damage derives from a breach of EU AI Act obligations, Article 18 presumes a causal link unless rebutted. Compliance with the EU AI Act alone does not exclude a defendant's liability.[7]

Suno consumer-terms investigation

On October 6, 2026, AGCM announced an investigation into Suno's potentially unfair consumer terms, including unilateral changes, copyright licences and U.S. arbitration. A public consultation is planned.[8] See October 7 digest coverage.

AI Act implementation and training decree

Legislative Decree No. 179/2026 was published October 8, 2026 and enters into force October 23. It allocates national supervision, provides enforcement and sandbox arrangements, and addresses AI training and final employment decisions.[9]

References