Tuscany AI Law: what it provides and what changes

Tuscany AI Law: with the final approval of the Regional Assembly, Tuscany becomes the first Italian Region to regulate artificial intelligence through a comprehensive law. At its core is the Tuscan Charter of Digital Rights: mandatory disclosure when interacting with an AI system, understandable explanations of algorithmic decisions, and the possibility to request human review in automated procedures. Alongside this, the law establishes a Regional Observatory on AI and strengthens the “Punti Digitale Facile” network. For local companies, the practical consequence is clear: showing how and why an AI system reaches a certain decision is no longer an optional best practice — it becomes an explicit expectation.

The Regional Assembly’s approval

Tuscany is the first Italian Region to adopt a comprehensive law on artificial intelligence. The Regional Assembly approved it definitively after a process lasting about five months: presentation to the Regional Government on February 16, 2026, approval by the Culture Committee on April 22, joint opinion from the Environment and Institutional Affairs Committees in mid-July, and finally the vote in the Assembly.

It is a measure worth reading carefully, because it does not merely state principles: it introduces operational tools, while also leaving open technical and political issues that are important to understand — especially for those, like Tuscan businesses, who will have to deal with these rules in practice.

What the Tuscan Charter of Digital Rights provides

The core of the measure is the Tuscan Charter of Digital Rights, which grants citizens:

• the right to be informed when interacting with an artificial intelligence system;

• the right to receive understandable explanations about algorithmic decisions concerning them;

• the right to request human intervention in the case of automated administrative decisions.

These principles reflect the approach already present in the GDPR and the European AI Act, but here they are implemented at a regional level and connected to concrete operational tools: a Regional Observatory on Artificial Intelligence, developed with the technical-scientific support of IRPET, and the strengthening of the Punti Digitale Facile network for digital literacy across the territory. The Metis Consortium is identified as the operational arm for the digitalization of local authorities and healthcare.

A less expected and so far little-discussed element is the recognition — alongside measures for the protection of minors and the right to disconnect — of what has been defined as the right to “digital happiness”: the protection of emotional well-being and mental health in a hyperconnected ecosystem. It is an unprecedented formulation in the Italian regulatory landscape, and its practical application will likely be one of the most delicate aspects to observe in the coming months.

Why the law has been criticized: the issue of regulatory sandboxes

The measure did not receive unanimous support: during the Assembly debate, criticisms were raised — particularly by the opposition — regarding the absence of tools for preventive risk assessment and regulatory sandboxes, controlled experimentation environments where companies could test AI solutions within a clear regulatory perimeter. This is likely the most concrete point raised in the debate, because it directly concerns the procedural clarity businesses need when adopting AI in their processes.

Does Tuscany have the infrastructure to guarantee these digital rights?

One objection raised in the Assembly deserves attention because it touches on a real issue: according to the debate, about 40% of Tuscan territory still falls within the so-called “white zones,” with insufficient internet coverage. The paradox highlighted by the opposition is that a law recognizing advanced digital rights — algorithmic explainability, disconnection, digital happiness — coexists with a still unresolved basic problem: access to connectivity itself.

It is a useful reminder: the digital transformation of Tuscan businesses and territories requires not only an advanced regulatory framework but also solid basic infrastructure. These two dimensions — rules and infrastructure — must progress together for the law to produce real effects across the entire territory, including underserved areas where many manufacturing SMEs operate and could greatly benefit from adopting AI.

What role does Tuscany aim to play in public AI research?

The law fits into a long-standing trajectory. Tuscany was already the first Region to adopt the principles of the AI Act through Regional Law 57/2024 on digital innovation. Added to this is participation in the interregional project Reg4IA, together with Emilia-Romagna, Friuli Venezia Giulia, and Sardinia, focused on territorial resilience and generative AI applied, among other things, to Civil Protection emergency management — a point highlighted by Councillor Alberto Lenzi in response to those who, citing the Liguria case (which applied AI to reduce healthcare waiting lists), accused Tuscany of adopting a more declarative than operational approach.

There is also a more ambitious dimension, less visible in official communication but relevant for understanding long-term strategy: following the Manifesto presented in Bologna on October 9, 2025, the Region supports the candidacy to host a node of a European AI Research Center, informally described as the “AI CERN,” with the declared goal of developing public research independent from major private tech players. Tuscany’s candidacy relies on concrete scientific assets: a strong university system, the presence of the CNR, the Center for Big Data and Artificial Intelligence, and the international experience gained through the Virgo project in gravitational wave physics.

What the law provides for employment and professional reskilling

On the employment front, the law focuses on professional reskilling through the Regional Employment Agency and Job Centers, with the declared goal of turning AI from a threat into a lever for work enhancement. This is where political debate becomes more heated, because it touches on a real and widespread concern among workers and businesses: not whether AI will change certain professions — it is already clear that it will — but whether institutions will be able to support this change with concrete tools, not just statements of intent.

Why this debate concerns businesses, not just politics

Beyond political controversy, the Tuscan law anticipates an issue that every company adopting AI will have to face, regardless of the regional regulatory framework: the ability to explain, trace, and govern decisions made with the support of automated systems. This is not an obligation born with this law — it is already implicit in the European AI Act — but the Tuscan law makes it visible and brings it into the local public debate.

There is also a more practical question that this type of regulation brings to the surface, and many companies are already asking it independently of the Tuscan law: are general-purpose AI assistants already available in companies (Microsoft Copilot 365 being the most widespread example) sufficient on their own to guarantee the required traceability and governance, or is something more needed? The answer depends on the use case. For individual productivity tasks, a well-configured general-purpose tool may be sufficient. When AI enters critical production or documentation processes — such as quality control, HR management, or regulated technical support — the focus shifts to native integration with existing company systems (ERP, MES, sector-specific management systems) and traceability of individual decisions, not just conversations: features that general-purpose assistants offer more limitedly compared to solutions specifically designed for that context.

From this perspective, the experience of those who design and implement AI systems in regulated industrial environments offers a useful contribution to the debate. At Aidia, we work daily with companies that must ensure traceability and transparency of their AI systems to comply with sector regulations — from manufacturing to pharmaceuticals — and this need, regardless of the Tuscan law, is already reshaping how AI is designed and adopted in companies. These cases differ from the scope of the Tuscan law — they arise from sector regulations, not from a regional AI law — but they share the same underlying principle that the Tuscan Charter of Digital Rights now brings into public debate: an artificial intelligence system that decides, signals, or responds must always be able to account for how it reached that decision.

Sources:

• Toscana Notizie, Official press release of the Tuscany Region “Ethical AI, Tuscany approves first comprehensive law in Italy: Giani and Lenzi, ‘Model for the Country’”, (21 July 2026) https://www.toscana-notizie.it/-/ia-etica-toscana-approva-prima-legge-organica-in-italia-giani-e-lenzi-un-modello-per-il-paese-

• InToscana, “Ethical AI: Tuscany approves the first law in Italy. Giani and Lenzi: ‘Model for the Country’”, (22 July 2026) https://www.intoscana.it/it/ia-etica-legge-toscana/