Italy’s Landmark AI Law: What It Does and Why the UK Should Pay Attention
Italy has made history by becoming the first EU member state to pass a national artificial intelligence (AI) law, which received final parliamentary approval on September 17, 2025. This significant legislation complements the EU AI Act and emerges from Italy’s “National AI Strategy,” initiated back in 2020, which emphasized the necessity for an ethical regulatory framework to ensure transparency, accountability, and reliability in AI systems, thereby fostering citizens’ trust and engagement.
Key Objectives of the Law
The newly enacted law aims to establish human-centric guardrails around the deployment of AI technologies while simultaneously promoting innovation across various sectors of the economy. By integrating a mix of sectoral rules, updates to criminal law, clarifications on copyright, and institutional choices, Italy provides a framework that could serve as a blueprint for other countries, including the UK, as they navigate their own regulatory pathways.
Summary of Key Provisions
- Human Oversight and Traceability: The law mandates that AI-assisted decisions must remain under human oversight. For instance, in the healthcare sector, medical decisions like prevention, diagnosis, and treatment must be made by healthcare professionals rather than AI systems. Patients must be informed when AI technologies are utilized in their care.
- Access by Minors: Children under the age of 14 are permitted to access AI technologies only with parental consent.
- Criminal Penalties: The law introduces a new offense concerning the unlawful dissemination of AI-generated or manipulated content, such as deepfakes, imposing prison terms of one to five years for causing unjust harm. Additionally, it increases penalties for using AI in existing crimes, such as market manipulation.
- Governance and Enforcement: The enforcement of the law is entrusted to two existing governmental bodies: the Agency for Digital Italy (AgID) and the National Cybersecurity Agency (ACN). Furthermore, the Department for Digital Transformation will guide the national AI strategy.
- Investment Signal: Up to €1 billion has been allocated to support companies operating in AI, cybersecurity, quantum technologies, and telecommunications.
- Copyright:
- Copyright Protection of AI Works: The law stipulates that works created with “AI assistance” are eligible for copyright protection if they are the result of “genuine human intellectual effort”. However, it leaves open the question of whether works solely produced by AI will attract copyright protection.
- Text-and-Data Mining (TDM): The legislation allows for text extractions or reproductions from works for AI model training, provided the principles of the DSM Directive are respected, including copyright owner opt-outs and lawful access to works. This provision potentially extends beyond the EU AI Act’s intended scope.
Implications for the UK
Italy’s proactive approach to AI regulation suggests that member states are willing to advance national legislation rather than await further EU-level adjustments. The Italian law augments the EU AI Act by introducing additional operational duties and specific criminal offenses, contributing to the EU’s complex regulatory landscape.
UK regulators have historically favored a more flexible, pro-innovation approach and are likely to maintain a “wait and see” stance on AI regulation. While Italy’s law alone may not compel the UK to pivot, a proliferation of similar national measures across the EU could pressure the UK to align its regulations, particularly concerning criminal sanctions and protections for minors.
Furthermore, the UK government’s previous proposals to extend the TDM exception were shelved after significant opposition from creative industries. As the current government continues to deliberate on copyright and AI, Italy’s legislation will likely influence future decisions on this critical issue.
As companies offering AI solutions in the EU prepare for compliance with country-specific obligations ahead of key AI Act deadlines, adopting an “EU-plus” baseline approach may help mitigate fragmentation and enforcement risks.
