The Meloni government on Wednesday approved two draft decrees implementing Italy’s AI law across education, labor, justice, public administration and law enforcement. But beyond the technical provisions, officials framed the initiative as an attempt to ensure that artificial intelligence remains subordinate to human judgment, responsibility and fundamental rights.
Why it matters: Italy is seeking to position itself at the forefront of AI governance in Europe while remaining aligned with the EU AI Act.
- The rules touch some of the most sensitive areas of AI deployment, from hiring decisions and workplace rights to biometric surveillance and criminal investigations.
- Rome is presenting AI not only as an economic challenge but as an ethical and societal one.
- The government’s emphasis on human oversight runs through every major provision of the package.
The Vatican connection: The most distinctive feature of the Italian approach is the philosophical framework underpinning it.
- In the official explanation accompanying the decrees, Palazzo Chigi says the measures are built around an “anthropocentric” vision in which AI can support decisions, services, education and competitiveness, but cannot replace human responsibility or limit fundamental rights.
- The government explicitly links that approach to Pope Leo XIV’s Magnifica Humanitas, arguing that both share a common principle: technology cannot become “the measure of the human being” nor replace human conscience, responsibility and judgment.
- AI, according to the government’s formulation, is valuable only when governed through an ethical and humanistic vision capable of combining innovation, justice, security and the common good.
- For Rome, the reference serves as the organizing principle of the entire package.
The big picture: The government argues that the decrees complete the national implementation of the EU AI Act rather than create a parallel regulatory framework. Officials say the measures were developed in continuous dialogue with the European Commission and are designed to fill areas that remain under national competence.
- The package now moves to parliamentary committees, regional authorities and independent regulators before final approval.
- Italian officials portray the effort as an attempt to reconcile innovation and safeguards: expanding AI adoption while protecting workers, citizens and consumers from opaque or fully automated decision-making.
Zoom in: AI enters schools and universities. Education is one of the pillars of the strategy. AI will be integrated into school curricula, civic education programs and teacher training initiatives, with a particular emphasis on understanding both the opportunities and risks associated with the technology.
- The government also plans broader AI literacy initiatives for adults, universities, technical institutes, public servants and healthcare professionals, arguing that widespread competence is a prerequisite for safe adoption.
- Officials view training as a national competitiveness issue as much as an educational one.
Workers get protection from algorithmic decisions. One of the most consequential provisions concerns employment. The draft legislation states that hiring, contract changes, disciplinary measures and dismissals cannot be based exclusively on automated decision-making systems. Final decisions must remain in the hands of a human decision-maker.
- Employees will also have the right to request an understandable explanation of decisions affecting them, including the role played by AI systems in the process. A dismissal based solely on an automated system would be considered null.
- The measures reflect a broader effort to prevent discrimination, opacity and excessive reliance on algorithms in workplace management.
Security, but no “Big Brother” The government’s security provisions attempt to strike a balance between technological capability and civil liberties.
- AI-powered biometric identification would be permitted only in exceptional circumstances, such as preventing serious threats to public safety, locating missing persons or identifying suspects linked to serious crimes. Any use would require judicial authorization and be subject to strict time, geographic and procedural limits.
- The decrees also prohibit biometric databases built through indiscriminate collection of online data and explicitly reject forms of generalized surveillance.
- Throughout the text, the government repeats the same principle: AI may provide analysis and support, but decisions with legal consequences must remain under human control.
Justice and accountability. The package also addresses liability and access to justice.
- Victims of harm caused by AI systems would receive stronger procedural protections, including access to technical documentation and mechanisms designed to reduce information asymmetries between individuals and technology providers.
- On the criminal side, the government plans to introduce a new offense targeting the omission or manipulation of safety measures in high-risk AI systems when those actions create a concrete danger to public safety or national security.
- The underlying message is consistent with the broader philosophy of the package: responsibility remains human, even when decisions involve advanced technologies.
Follow the money. Regulation is only one side of Rome’s AI strategy.
- The government has earmarked up to €1 billion from its venture capital support framework to strengthen Italy’s AI ecosystem and support domestic technological development. According to official figures, Italy’s AI market reached €1.8 billion in 2025, growing 50% year-on-year.
- Officials say the goal is to build national capabilities in strategic sectors while reducing dependence on foreign technologies and strengthening Europe’s broader digital sovereignty agenda.
The bottom line: Italy’s AI package is about more than regulatory compliance with Brussels.
- The Meloni government is trying to advance a distinctly human-centered model of technological governance — one that explicitly draws on the ethical language of Magnifica Humanitas and seeks to ensure that, even as AI spreads through schools, workplaces, courts and security agencies, the final responsibility remains with people rather than machines.
