Peri News / Italy
Italian Supreme Court report examines AI policing and liability
Diritto.it describes a non-binding report on Italian rules for investigative AI and compensation. Its account and application dates require verification against official texts.

Diritto.it, in its article “Intelligenza artificiale: la relazione della Cassazione”, reports that the Italian Supreme Court of Cassation’s research and case-law office examined Legislative Decree No. 160 of 9 September 2026 in Report No. 63 of 11 November 2026. The decree is described as adapting Italian law to the European Union’s AI Act. The report is a non-binding analysis, rather than a judgment. Neither the report nor the decree has been consulted directly: the account requires verification against the official texts, including the provisions’ application dates, which the article does not specify.
For investigations in Italy, Diritto.it describes a new Article 359-ter of the Code of Criminal Procedure governing identification and location through real-time remote biometric systems. Their use would ordinarily require authorisation from the preliminary investigations judge, with specific procedures for urgent cases; breaches of statutory conditions could render the results inadmissible. The account highlights safeguards for biometric data and human control over investigative assessments. It also attributes new offences concerning omitted safety measures and unlawful alteration of AI systems to Article 437-bis of the Criminal Code, additionally mentioning failure to provide the human oversight required for certain high-risk systems.
On compensation, according to Diritto.it, Articles 16–20 would introduce mechanisms for accessing evidence, presumptions concerning causation and, in specified cases, a direct claim against the insurer. However, the analysis as reported rules out a separate general civil liability regime for AI: the contractual and non-contractual liability rules relevant to each dispute would continue to apply. The main difficulties concern access to technical information and attributing conduct to developers, providers, users and those responsible for oversight. Compliance with the AI Act would not, by itself, provide an automatic defence to civil liability.
Sources
- Intelligenza artificiale: la relazione della Cassazionediritto.it — 20 November 2026
