Introduction
Italy's Law No. 132/2025, effective October 10, 2025, marks a significant milestone as the first national AI legislation within the EU. This law complements the existing Regulation (EU) 2024/1689 (EU AI Act), by addressing areas not covered by EU regulation, establishing a precedent for national AI frameworks in Europe. The legislation introduces clear governance structures and sector-specific safeguards for critical areas such as healthcare, employment, justice, and intellectual professions. Furthermore, it enhances protections for minors, democratic institutions, and intellectual property rights in the context of artificial intelligence. This article provides a comprehensive overview of these transformative changes.
Relationship to the EU AI Act
A critical feature of Italy's AI Law is its complementary relationship with the EU AI Act. The Italian national law must be interpreted and applied in accordance with EU AI Act rules and definitions, ensuring consistency across the European regulatory landscape.
Rather than creating parallel or conflicting requirements, the Italian law fills specific gaps in the EU framework with Italy-specific provisions addressing national priorities in sectors like healthcare, employment, justice, and culture. Importantly, the law does not introduce new compliance obligations beyond the EU AI Act; instead, it provides additional sectoral guidance and safeguards within the boundaries established by European legislation.
Definitions
Article 2 of Italy’s AI Law adopts the definitions set out in the EU AI Act, ensuring full alignment with the European framework. Most key terms, including artificial intelligence system and artificial intelligence model, follow the EU definitions directly.
The only definition introduced at the national level is “data,” defined as any digital representation of acts, facts, or information, including sound, visual, or audiovisual data. For all other terminology, the law refers to the corresponding provisions of the EU AI Act.
Structure
Italy's AI law is divided into 6 chapters comprising 28 articles in total.
- Chapter I: General Principles: Establishes fundamental principles, national AI strategy framework, and guiding values
- Chapter II: Sector-Specific Rules: Contains provisions for healthcare, scientific research, employment, justice, and intellectual professions
- Chapter III: Governance and Authorities: Defines the roles of AgID, ACN, and other regulatory bodies
- Chapter IV: Protection of Users and Copyright: Addresses data protection for minors, intellectual property rights, and AI-assisted works
- Chapter V: Criminal and Procedural Law: Introduces new offenses and procedural modifications for AI-related disputes
- Chapter VI: Final Provisions: Covers implementation timelines, delegated powers, and international cooperation
Principles
Article 4 of the Act establishes the foundational principles that govern AI deployment in Italy. These principles reflect a human-centric approach to AI governance and include:
- Data Protection and Privacy: The law mandates lawful, correct, and transparent processing of personal data, reinforcing Italy's commitment to GDPR principles in the AI context.
- Fundamental Rights Centrality: The law places fundamental rights at the center of AI governance, ensuring that technological advancement does not come at the expense of constitutional protections and human dignity.
- Safety and Transparency: AI systems must be safe for users and stakeholders, with clear transparency requirements that enable individuals to understand when and how AI affects them.
- Proportionality: Regulatory requirements must be proportionate to the risks posed by AI systems, avoiding unnecessary burdens while ensuring adequate protection.
National AI Strategy
Beyond specific regulatory requirements, the Act establishes a framework for ongoing strategic planning. The National AI Strategy must be updated every two years by the Interministerial Committee for Digital Transition, ensuring that Italy's approach evolves with technological developments. This strategy serves as a reference framework for policy and regulatory decisions across government, with support from the Department for Digital Transformation. The regular update requirement ensures that Italy's AI governance remains responsive to emerging challenges and opportunities.
Governance Structure
One of the most significant features of Italy's AI Law is its governance architecture, which distributes regulatory responsibilities between two primary authorities:
- Agency for Digital Italy (AgID): AgID serves as the notifying authority responsible for innovation promotion and conformity assessment. This positioning reflects Italy's commitment to fostering AI innovation while ensuring technical compliance with regulatory standards.
- National Cybersecurity Agency (ACN): ACN functions as the market surveillance authority responsible for oversight, inspection, and sanctions. Additionally, ACN serves as Italy's single point of contact with EU institutions on AI matters, streamlining international coordination.
This dual authority model balances the objectives of promoting innovation and ensuring compliance, with clear delineation of responsibilities to avoid regulatory overlap or confusion.
Coordination Mechanisms
Recognizing that AI impacts multiple regulatory domains, the law establishes coordination mechanisms. A Coordination Committee at the Presidency of the Council of Ministers provides high-level oversight and ensures consistency across government agencies.
The law preserves existing powers for specialized authorities, including the Garante (Italian Data Protection Authority) for privacy matters and AGCOM for communications. It also mandates collaboration with sectoral regulators, including the Bank of Italy, CONSOB (securities regulator), and IVASS (insurance regulator), ensuring that AI oversight leverages existing regulatory expertise. This coordinated approach prevents regulatory fragmentation while respecting the specialized knowledge of sector-specific authorities.
Sector-Specific Provisions
1. Healthcare & Scientific Research
Italy's AI Law includes particularly innovative provisions for healthcare and scientific research (articles 7-8), recognizing both the transformative potential of AI in these fields and the sensitivity of health data.
- Data Processing for Research: One of the most significant provisions allows secondary use of personal data (including sensitive health data) for public interest scientific research without requiring new consent. This provision aims to accelerate medical AI development while maintaining privacy protections through mandatory removal of direct identifiers. Before processing begins, organizations must provide 30-day notification to the Garante, including details of GDPR compliance measures (article 8). This secondary use is identified as serving a significant public interest under the Italian Constitution, reflecting Italy's commitment to advancing medical research.
- AI in Clinical Settings: The law establishes clear boundaries for AI in healthcare delivery. AI must function only as a support tool and cannot discriminate against patients or independently decide access to treatment. Final medical decisions must remain with human healthcare professionals, preserving the physician-patient relationship and professional responsibility.
2. Employment
The use of AI in the workplace (Article 11) is mandated to improve working conditions, protect the psychophysical integrity of workers, and increase productivity.
- Principles: AI use must be safe, reliable, transparent, and cannot conflict with human dignity or violate the confidentiality of personal data.
- Non-Discrimination: AI in managing the employment relationship must guarantee compliance with the inviolable rights of the worker without discrimination (including based on sex, age, ethnic origin, etc.).
- Information Obligation: The employer or the client is required to inform the worker of the use of artificial intelligence.
- Monitoring: The Observatory on the adoption of artificial intelligence systems in the world of work is established at the Ministry of Labor and Social Policies. This body is tasked with defining a strategy for AI use in the workplace, monitoring its impact on the labor market, and promoting the training of workers and employers in the field of artificial intelligence.
- Training Delegation: The Government is delegated to adopt decrees that provide for literacy and training paths on AI systems for professionals and workers.
3. Justice System
The Italian AI Law introduces specific safeguards for the use of artificial intelligence within the justice system, reaffirming that human judgment remains central to all judicial activity.
- Reserved Decision-Making: Decisions involving the interpretation and application of law, the evaluation of facts and evidence, and the adoption of judicial measures are strictly reserved for magistrates. AI may support, but never replace, the exercise of judicial discretion or reasoning.
- Authorized Uses and Oversight: AI systems may be used only for support functions, such as the organization of justice services, simplification of judicial work, and ancillary administrative activities. The Ministry of Justice regulates these applications and authorizes experimental use of AI in judicial offices, pending full implementation of the EU AI Act, after consulting the designated national authorities.
- Judicial Training: The Ministry of Justice is tasked with promoting training and awareness programs for judges and administrative staff on AI technologies. These initiatives focus on digital literacy, responsible use, and understanding of AI-related risks and benefits within judicial operations.
- Jurisdiction for AI Disputes: The law delegates to the Government the power to define rules governing the use of data, algorithms, and mathematical models in AI systems. Disputes relating to AI operation fall under the jurisdiction of specialized business sections, with corresponding amendments introduced to the Code of Civil Procedure.
4. Intellectual Professions
The law addresses how AI affects traditional intellectual professions such as law, accounting, architecture, and engineering.
- Protecting the Fiduciary Relationship: Professionals must inform clients about the use of AI systems in delivering intellectual services, maintaining trust and awareness of AI involvement.
- Communication Standards: Information must be presented in clear, simple, and comprehensive language. AI systems act as support tools, assisting professionals in law, engineering, accounting, and similar fields, but cannot replace the professional’s final judgment, and clients must be fully informed of their use.
5. Democratic Institutions
The law explicitly prohibits AI use that could interfere with democratic institutions or distort public debate. This includes protections against disinformation campaigns, opinion manipulation, and AI-generated content designed to undermine electoral integrity or public discourse.
Data Protection Provisions
1. Minors Protection
The Act establishes age-differentiated consent requirements for minors' data in AI systems:
- Young Children: Parental consent is required for children under 14, recognizing their developmental limitations in understanding AI's implications.
- Adolescents: Children aged 14-18 can provide their own consent if information about AI systems is accessible and comprehensible to them. This graduated approach respects developing autonomy while ensuring age-appropriate protection.
2. Healthcare Data Processing
To promote innovation while protecting privacy, the law provides a framework for processing health data in AI development:
- Public Interest Declaration: Processing personal and sensitive data for AI-driven scientific research (e.g., prevention, diagnosis, drug development, public health) is recognized as serving the public interest, consistent with Articles 32 and 33 of the Italian Constitution and GDPR Article 9(2)(g).
- Secondary Use Authorization: Secondary use of personal data without direct identifiers is permitted for research, even for sensitive categories, without additional consent when initial consent is legally provided.
- Transparency: Data controllers may fulfill information obligations via general disclosures on their websites.
3. Sports Research
The law also permits processing data related to athletic performance for research purposes:
- Permitted Processing: Data may be anonymized, pseudonymized, or synthesized for studying athletic gestures, movements, and performances across all sports.
- Condition: Such processing requires appropriate information to the data subject.
Intellectual Property
The Italian AI Law updates copyright legislation (Law 22 April 1941, n. 633) to clarify the status of works created with AI assistance and to regulate the use of copyrighted material for AI training.
- Copyright Protection for Human-Created Works: The law reinforces that copyright applies only to works originating from human creativity:
- Protection is limited to works reflecting the author’s intellectual effort.
- AI-generated works without meaningful human contribution are not eligible for copyright, distinguishing them from human-created works.
- Text and Data Mining (TDM) for AI Training: The law permits the use of copyrighted material to train AI models, under specific conditions:
- Reproduction or extraction of works from databases or networks to which one has legitimate access is allowed for AI text and data mining, including generative AI.
- TDM must respect Articles 70-ter and 70-quater of the Copyright Law, which allow content owners to exercise opt-out rights.
- Violations of these provisions, including unauthorized reproduction or extraction through AI systems, are subject to penalties under the Copyright Law.
Criminal Law Provisions
The Act creates new criminal offenses specifically addressing AI-related harms:
- Deepfake Dissemination: A new offense criminalizes the unlawful dissemination of AI-manipulated content (deepfakes) with imprisonment penalties. This addresses growing concerns about synthetic media used for fraud, defamation, or manipulation.
- AI as Aggravating Factor: The law establishes a general aggravating circumstance for crimes committed using AI tools, recognizing that AI can amplify criminal harm through scale, sophistication, or targeting capabilities.
The law amends Italy's Code of Civil Procedure to address AI-related disputes, ensuring that judicial processes can effectively handle novel questions arising from AI deployment.
Compliance Considerations
Organizations deploying AI in Italy must navigate a multi-layered compliance landscape:
- EU AI Act Baseline: All EU AI Act obligations apply in Italy, including risk classification, conformity assessment, transparency requirements, and record-keeping for high-risk systems.
- Sector-Specific Requirements: Additional obligations apply depending on the sector and use case. Healthcare organizations, employers, justice system entities, intellectual professionals, and others must comply with sector-specific provisions described above.
- Enhanced Transparency: Beyond EU requirements, certain applications require broader information disclosure, particularly in employment contexts.
- Impact Assessments: Organizations must conduct assessments to prevent algorithmic discrimination and ensure fundamental rights protection, particularly when deploying AI in sensitive contexts.
Interaction with Existing Frameworks
Italy's AI Law builds on and integrates with existing regulatory frameworks rather than replacing them:
- Labor Law: AI employment provisions build on Italy's existing labor protection laws and rules governing remote monitoring of workers.
- GDPR: Data protection requirements integrate with existing GDPR obligations, including data protection impact assessments (DPIAs) and privacy by design principles.
- Copyright Law: Intellectual property provisions leverage and extend existing Italian copyright law frameworks.
International Context
Italy's pioneering legislation has significant implications beyond its borders:
- EU Precedent: As the first EU member state with comprehensive national AI legislation, Italy provides a model for how other member states might supplement the EU AI Act with national provisions.
- National Sovereignty: The law demonstrates how member states can address national priorities and concerns within the boundaries of EU harmonization, balancing centralized European regulation with local democratic control.
- Regulatory Innovation: Italy's provisions on copyright for AI-assisted works, healthcare data for research, and criminal liability for deepfakes may influence other jurisdictions grappling with similar questions.
Conclusion
Italy's Law No. 132/2025 represents a landmark development in AI governance, the first comprehensive national AI law among EU member states. By complementing the EU AI Act with targeted sectoral provisions, Italy has charted a path for how member states can address national priorities while maintaining European regulatory harmonization.
The law's sector-specific approach to healthcare, employment, justice, and intellectual professions provides concrete guidance for high-stakes AI applications while preserving fundamental principles of human dignity, professional responsibility, and democratic integrity. Its innovative provisions on copyright protection for AI-assisted works and criminal liability for deepfakes address emerging challenges at the frontier of AI law.
For organizations operating in Italy, the law creates both obligations and opportunities. While compliance requires attention to sector-specific requirements beyond the EU AI Act baseline, the law also provides legal certainty and pathways for innovation, particularly in healthcare research and AI-assisted creative work.
As implementation proceeds and other member states observe Italy's experience, Law No. 132/2025 may prove influential far beyond Italian borders, shaping how Europe balances AI innovation with fundamental rights protection in the years ahead.
Organizations should view compliance not as a one-time exercise but as an ongoing process of monitoring regulatory developments, engaging with authorities, and adapting practices as Italy's AI governance framework matures. With thoughtful implementation, Italy's AI Law can support responsible innovation that serves both economic competitiveness and societal values.
