In a recent ruling, the Lombardia ‘TAR’ (Regional Administrative Court) reported to the competent Bar Association a lawyer who, during the trial, used artificial intelligence (AI) systems in a manner contrary to the duty of loyalty and integrity.
In his appeal, the lawyer had cited irrelevant case law, with extracts that in many cases referred to unknown jurisprudential trends.
The case is an example of the risk of hallucinations presented by generative AI tools, which, in a nutshell, consist in the generation of incorrect content (here, represented by judgements that are irrelevant to the case in question and trends that are, in all likelihood, non-existent).
In the opinion of the TAR, the lawyer's conduct constitutes a violation of the duty of the defence to behave with loyalty and integrity in court, such as to introduce elements that are potentially capable of influencing the fair trial and the decision-making phase in an incorrect manner, as well as making the review of the case law cited and the principles expressed by both the judge and the opposing parties unnecessarily burdensome.
The lawyer's statement that he had cited case law found using AI-based search tools that generated incorrect results was considered by the TAR to be without exonerating value, given that the signing of procedural documents aims to attribute responsibility for the outcome of the defence briefs to the signatory, regardless of whether they drafted them personally, using AI tools or with the help of their colleagues.
The judge also referred to the principle of the centrality of human decision, established in the “Carta dei principi per un uso consapevole dei sistemi di intelligenza artificiale in ambito forense” drawn up by the Milan Bar Association in 2024, according to which professionals have a duty to verify and check the results of searches carried out using AI systems.
The decision is consistent with the provisions of the recent Italian law on AI in relation to intellectual professions. Article 13 of Law No. 132/2025 stipulates that the intellectual work performed by the professional must prevail over the use of AI systems. The latter, in any case, is only permitted if it is used solely for the purpose of carrying out activities that are instrumental and supportive of the professional activity.
Articoli Correlati