Page language: Italiano English Français Deutsch

Artificial intelligence laws around the world

By the FAIND newsroom · Updated on 17 October 2026

A hand resting on an ornate globe, with a judge's gavel in the foreground

There is no worldwide law on artificial intelligence. There are different approaches: Europe regulates according to risk, the United States relies on the market, China on control. Here is the map.

In brief

The European Union has the world's first comprehensive law: the AI Act.
Italy has its own national law alongside the European one.
The United States and China follow very different paths.

European Union: the AI Act

The AI Act has been in force since 2024 and applies in stages. It classifies AI systems by risk: unacceptable (banned, such as social scoring of citizens), high (allowed with strict obligations, for example in healthcare, employment, education and justice), limited (transparency obligations, such as declaring that you are talking to a chatbot or that content is artificial) and minimal (no obligations). Specific rules cover large general-purpose models.

Italy

In 2025 Italy passed a national law on artificial intelligence that sits alongside the European regulation: among other things it covers healthcare, work, public administration, justice and copyright, and introduces criminal provisions against harmful deepfakes. The authorities in charge are the Agency for Digital Italy and the National Cybersecurity Agency.

United States

There is no comprehensive federal law. Rules come from presidential executive orders, which change with each administration, from sector agencies and from individual states, some of which have passed laws of their own. The general approach favours innovation and competition with China.

China

China has introduced targeted rules for individual areas: recommendation algorithms, synthetic content and generative AI services, with obligations to register and to label generated content. State control is more direct than elsewhere.

The United Kingdom and other countries

The United Kingdom has so far chosen not to pass a single law, leaving the rules to sector regulators. Japan and South Korea have passed framework laws aimed at promoting development. Internationally there are shared principles, such as those of the OECD, and a Council of Europe convention on AI and human rights, but no treaty that is binding on everyone.

Why it concerns you too

European rules apply to anyone offering AI services in the Union, even if the company is based elsewhere. This is why some features arrive in Europe later. News on laws and rules is gathered every hour on the FAIND home page, in the “Law & policy” section.