Source: IUS In Itinere

The progressive digitalization of public activity and the introduction of artificial intelligence in courtrooms require renewed reflection on constitutional principles governing judicial function, including reservation of jurisdiction, the right to a fair trial, protection of fundamental rights, and judge impartiality and independence. The EU Regulation on Artificial Intelligence n. 1689/2024 classifies judicial AI systems as “high-risk” and frames their permissible use, asserting that final decision-making must remain human-led; this interpretation is reinforced by the Regulation’s considering and by its large-scope formulation regarding AI assistance to judicial authorities. At the domestic level, Law n.132/2025 transposes the Regulation and addresses the use of AI in judicial activity, notably in articles 14 and 15—permitting organizational and administrative simplifications—and article 17, which expands exclusive tribunal competence to disputes concerning the functioning of AI systems; the author notes ambiguous definitions in both EU and national texts but regards them as a necessary starting point.

The article interrogates whether the judge’s role might be compromised when receiving and evaluating digital evidence (“digital evidence”), focusing on civil-procedure difficulties in mapping the Regulation’s notion of “fact” to the “fatto notorio” of art. 115 cod. proc. civ.; jurisprudence traditionally values widely shared information from press and broadcasting but does not automatically equate web information with notoriety, requiring verification of truth and source reliability. The piece surveys case law: Cass. Civ. 18748/2010 on press-based notoriety; a restrained Cass. Pen. decision (n. 36315/2016) deeming a Google Maps-derived document inadmissible when acquired unilaterally by a judge; more permissive rulings such as Cass. Civ. n.16165/2011, Cass. Pen. Sez. III n.35869/2023, and TAR Calabria n.1604/2018 that accepted Google Earth or aerofotogrammetries as evidentiary material; and mixed treatment of AI tools like ChatGPT (mere mention in Cass. n.14631/2024, and refusal of ChatGPT-generated queries in TAR Lazio n.4546/2025). The conclusion stresses potential advantages of digital evidence if adequate precautions are adopted and if informatized procedures do not become a means to evade core principles of the legal order.