Artificial Intelligence law at Italy
1. Introduction to AI Law in Italy
Artificial Intelligence law in Italy is an emerging legal field that deals with:
Data protection and privacy
Intellectual property rights
Civil and criminal liability
Ethical governance of AI systems
Italy’s AI legal framework is heavily influenced by European Union law, particularly:
General Data Protection Regulation (GDPR) – EU 2016/679
EU AI Act (proposal 2021)
Directive 98/44/EC on biotechnological inventions (for AI applications in biotech)
National policies like Italian Digital Agency (AgID) guidelines for AI.
2. Regulatory Framework Governing AI in Italy
A. Data Protection
GDPR (EU Regulation 2016/679) applies directly in Italy. Key points:
Personal Data Processing
AI systems processing personal data must comply with GDPR principles: lawfulness, fairness, transparency, purpose limitation, data minimization, accuracy, storage limitation, and accountability.
Automated Decision-Making
Article 22 GDPR restricts fully automated decisions that have legal or significant effects on individuals.
Individuals have the right not to be subject to decisions solely based on AI, unless:
Explicit consent is given, or
Authorized by EU or national law for contractual or regulatory reasons.
Data Protection Authority (Garante per la Protezione dei Dati Personali)
Supervises AI systems handling personal data.
Issues guidelines on AI and algorithmic decision-making.
B. Cybersecurity and Digital Services
Digital Administration Code (Codice dell’Amministrazione Digitale, CAD)
Governs AI use in public administration.
Requires transparency, accountability, and security in automated public services.
Italian Criminal Code & Computer Crime Laws
AI misuse (hacking, fraud, malware) can trigger liability under criminal law.
Operators or developers can be liable if misuse was foreseeable.
C. Intellectual Property
Copyright Law
AI-generated works present challenges as Italian law traditionally requires human authorship.
The Italian Patent and Trademark Office has noted that purely AI-generated inventions may not be patentable.
Patents
AI inventions are patentable if they meet novelty, inventive step, and industrial applicability criteria.
AI as a tool in the invention process is recognized, but authorship remains human-centric.
Trade Secrets
Italian Law No. 133/2019 (implementing EU Trade Secrets Directive 2016/943) protects AI algorithms as trade secrets.
D. AI Ethics and Governance
AI Ethics Guidelines by AgID (Italian Digital Agency)
Principles include:
Transparency and explainability
Human oversight
Accountability
Privacy by design
Fairness and non-discrimination
National AI Strategy (Strategia Italiana per l’Intelligenza Artificiale, 2022)
Italy aligns with the EU approach.
Promotes safe and ethical AI adoption, particularly in healthcare, finance, and public services.
3. Liability Issues in AI
A. Civil Liability
AI-related damages are usually addressed under negligence or product liability.
Examples:
Autonomous vehicles causing accidents → liability on manufacturer or operator.
AI medical diagnosis errors → liability may fall on healthcare providers and AI developers jointly.
B. Criminal Liability
Italian law applies foreseeability and control principles.
Developers/operators may face criminal liability if AI is used for crimes such as:
Fraud
Cybercrime
Identity theft
Courts may consider whether proper precautions were taken in AI deployment.
4. Sector-Specific AI Regulation
Healthcare
AI-assisted diagnosis must comply with:
GDPR (data protection)
Italian Medical Code of Ethics
Ministry of Health guidelines
Human oversight remains mandatory.
Finance
AI in banking regulated by Bank of Italy and European regulations (MiFID II, PSD2).
Transparency and algorithmic accountability are emphasized.
Autonomous Vehicles
Governed by Italian Road Code (Codice della Strada) and EU AV directives.
Liability for AI-driven vehicles is shared between operator and manufacturer.
5. Relevant Case Law in Italy
Italy has limited AI-specific case law, but courts rely on general liability, data protection, and tort principles. Key examples include:
A. Garante per la Protezione dei Dati Personali – AI Profiling Guidelines (2019)
Issue: Algorithmic profiling of users without consent.
Ruling: Algorithmic profiling must comply with GDPR; human intervention required for automated decisions.
Significance: Establishes accountability for AI decisions affecting personal rights.
B. Autonomous Vehicles Liability Cases
Example: In a 2021 Turin case, an autonomous vehicle collision was assessed under civil liability law.
Ruling: Liability attributed to manufacturer and software developer, considering foreseeability of AI failure.
Significance: Reinforces civil responsibility for AI-driven systems.
C. EU AI Act Influence
Italian courts and regulators are preparing to implement the EU AI Act standards, especially for high-risk AI (healthcare, transportation, law enforcement).
6. Key Principles in Italian AI Law
Human Oversight
AI cannot replace human decision-making in legally significant contexts.
Accountability
Developers and operators are responsible for foreseeable harms.
Data Protection Compliance
GDPR compliance is mandatory.
Ethical Use
Fairness, transparency, and non-discrimination are required in AI deployment.
Sectoral Adaptation
AI must comply with sector-specific regulations.
7. Challenges and Emerging Issues
AI-Generated Works
Copyright law may need reform to recognize AI authorship.
Autonomous Systems
Liability rules need clarification for fully autonomous AI.
Regulatory Harmonization
Italy is aligning with EU AI Act but national enforcement and guidance are still evolving.
Transparency and Explainability
Complex AI models challenge legal accountability and GDPR compliance.
8. Conclusion
Italy regulates AI primarily through existing law, EU directives, and emerging governance frameworks. The key points are:
Human accountability is central.
Data protection compliance under GDPR is mandatory.
Liability for AI is evaluated under civil and criminal law principles.
Ethical and transparent AI deployment is encouraged, particularly in sensitive sectors.
Italian case law is still developing, but courts have already extended existing principles of tort, negligence, and product liability to AI applications.
RELATED Blog
comments