Italy Is Leading Europe on AI Literacy — And Every Lawyer Should Pay Attention
Law 132/2025 establishes the most specific AI obligations for lawyers anywhere in Europe. The rest of the continent is watching.
On October 10, 2025, Italy became the first EU member state to enact comprehensive national AI legislation complementing the EU AI Act. Law No. 132/2025 goes beyond transposing EU requirements — it establishes sector-specific rules that exceed the EU baseline in several material respects, creating the most detailed set of AI obligations for legal professionals anywhere in Europe.
Italy has a distinctive regulatory temperament: enthusiasm for innovation combined with strong safeguards for professional integrity. Law 132/2025 is a product of that temperament, and legal professionals across Europe would be wise to study it closely.
The Mandatory Disclosure Requirement
The centrepiece of Law 132/2025 is its mandatory disclosure requirement. As A&O Shearman's analysis confirms, the law restricts regulated professions — including lawyers and accountants — to using AI only for "auxiliary tasks," and requires disclosure to clients of any AI use.
This is more prescriptive than anything the EU AI Act demands. Article 4 imposes a general literacy obligation. Law 132/2025 translates that into a concrete, enforceable rule: if you used AI, you must say so.
Consider the day-to-day work of a commercial lawyer. Drafting, research, document review, due diligence — each might involve AI assistance. Under Law 132/2025, any such use triggers a disclosure obligation. The lawyer must inform the client, regardless of whether the client asks. This creates a disclosure culture rather than a disclosure exception. Silence about AI use is not an option.
The "Auxiliary Tasks" Restriction
Equally significant is the restriction to auxiliary tasks. AI may assist with auxiliary functions; it may not replace professional judgment on substantive matters.
The legislation does not define "auxiliary" with granular precision, but the intent is clear. Tasks likely to qualify include legal research and preliminary analysis, document organisation and initial review, standard form generation, and initial identification of issues requiring human analysis. Tasks that cannot be delegated to AI include final case strategy decisions, substantive legal advice, negotiation of material terms, courtroom advocacy, and professional judgment calls on conflicts and ethics.
The distinction is practical. A lawyer who uses AI to draft an initial memorandum and then applies independent professional judgment to revise it is operating within the auxiliary framework. A lawyer who submits AI-generated advice without substantive independent review is not.
The Garante Precedent: Italy Means What It Says
Italy's willingness to enforce technology regulation is not theoretical. In March 2023, the Garante per la Protezione dei Dati Personali temporarily banned ChatGPT over data protection concerns. The ban was lifted after OpenAI implemented compliance measures, but the message was unambiguous: Italian regulators will act swiftly.
The Garante's intervention showed that Italian authorities possess the technical capacity to evaluate AI systems, the willingness to take commercially significant enforcement action, and a view of AI governance as a fundamental rights matter. For legal professionals, the combination of specific rules under Law 132/2025, established enforcement authority, and demonstrated regulatory willingness creates genuine compliance pressure. This is not a jurisdiction where guidance can be treated as aspirational.
AgID and Governance Infrastructure
Law 132/2025 establishes clear governance frameworks through the Agenzia per l'Italia Digitale (AgID), providing defined institutional pathways for compliance questions, reporting, and dispute resolution. Where other jurisdictions are still building administrative infrastructure, Italy has resolved institutional uncertainty from the outset.
Multi-Jurisdictional Implications
Italy's approach has consequences well beyond its borders. An international firm with offices in Milan, Frankfurt, and Amsterdam must navigate varying disclosure obligations across jurisdictions. The practical reality: on any matter involving Italian counsel, the safest approach is to adopt Italian disclosure standards across the entire team.
This is the first-mover effect in regulation. Italy's rules do not technically bind lawyers elsewhere, but the most prescriptive standard tends to become the de facto standard for entire matters. Italian disclosure norms will influence practice far beyond Italy itself.
What This Means for the Profession
Italy's approach deserves both admiration and careful analysis. The admiration is warranted because Law 132/2025 takes AI regulation seriously as a matter of professional integrity, not merely compliance. It treats lawyers as professionals who owe genuine duties of transparency, backed by enforceable rules.
The analysis is warranted because prescriptive regulation carries costs — the auxiliary-tasks restriction will require ongoing interpretation as AI evolves, and mandatory disclosure burdens will vary between a solo practitioner and a large firm managing complex transactions.
But Italy has demonstrated that it is possible to move from the EU AI Act's general principles to specific, workable professional obligations — before most other member states have begun the exercise. The rest of Europe is watching. Many will follow.
This article draws on research from the Twin Ladder Article 4 panoramic analysis, a comprehensive examination of the EU AI Act's literacy mandate and its implications for legal professionals across Europe.