Source: Il Sole 24 Ore

The Court of Cassation, in judgment number 29015/2026, held that evidence produced by IT tools based on artificial intelligence cannot be treated as newly discovered evidence sufficient to reopen a final conviction. The contested prosecution involved ideological forgery and fraud tied to an illegal over-invoicing scheme aimed at obtaining structural and EU funds for tourism in the Lombardy region, and the original conviction was based primarily on wiretaps indicating the defendant was a hidden partner in the beneficiary company. The defence sought review after presenting previously unpublished voice-analysis results from comparison software described as utilising artificial intelligence or comparative biometric analysis to argue the person on the recorded calls was not the appellant.

The Supreme Court acknowledged as a general principle that more advanced technologies can provide new objective data capable of supporting different technical assessments, but found the defence’s submission deficient. The Court emphasised that the defence did not explain the “new frontier” of the employed system, nor did it supply sufficient detail on methodologies, scientific reliability, or genuinely innovative features of the biometric comparison. The judges characterised the guidance on the AI-based tests as vague and lacking substantive methodology; the only certain fact was that an AI-based voice-identification programme had been used, which the Court deemed insufficient to overturn the appellate refusal to reopen the case.