AI regulation

In short

AI regulation is the set of laws, rules and official guidance that governments use to control how artificial intelligence is built and used. Approaches differ widely: the EU has one broad AI law, while the UK and US rely mostly on existing laws and regulators.

Also called: AI law, Regulation of artificial intelligence, AI governance law
Topic: Law, policy and the UK
Level: Beginner
First described: The best-known dedicated AI law is the EU's AI Act, Regulation (EU) 2024/1689, which came into force on 1 August 2024¹.

There is no single global AI rulebook. Each country decides whether to write a dedicated AI law or rely on laws it already has. The UK government, for example, points out that discriminatory AI outcomes may breach the Equality Act 2010 and that data protection law requires personal data to be processed fairly³.

The European Union chose a dedicated law. The EU AI Act, Regulation (EU) 2024/1689, entered into force on 1 August 2024 and applies in stages¹. It sorts AI by risk and puts strict duties on high-risk uses such as hiring and credit scoring². Its rules for general-purpose AI models have applied since 2 August 2025¹. In July 2026 the Digital Omnibus pushed the main high-risk deadlines back to December 2027 and August 2028¹.

The UK chose a lighter route. Its 2023 white paper set five principles, including safety, transparency, fairness, accountability and contestability, and said they would not be put into law at first but applied by existing regulators within their own areas³. As of October 2026 the only AI regulation bill before Parliament is a private member's bill from Lord Holmes of Richmond, which has not gone past first reading in the Lords since March 2025⁴.

The United States has no broad federal AI law. Policy is set largely by executive orders. In January 2025 President Trump revoked his predecessor's AI order, and a December 2025 order set up an AI Litigation Task Force to challenge state AI laws the administration sees as too burdensome, while calling on Congress for a single national standard⁵.

Voluntary standards fill some of the gaps. The US NIST AI Risk Management Framework, for instance, is designed to be voluntary and usable in any sector⁶.

Example

A UK company that uses AI to sift job applications is covered in Britain by the Equality Act 2010 and data protection law rather than an AI-specific statute³. If it sells the same tool into France, the EU AI Act treats it as high-risk².

Why it matters

The rules decide who is responsible when AI goes wrong and what rights you have when it is used on you. Because the EU, UK and US have chosen different paths, the same product can face very different duties depending on where it is sold.

Questions people ask

Does the UK have an AI law?

Not a dedicated one. The UK government's 2023 approach asks existing regulators to apply five AI principles within their own areas, without a new statute at first³. As of October 2026, the only AI regulation bill in Parliament is a private member's bill that has stalled at first reading⁴.

Is AI regulated in the US?

There is no broad federal AI law. The federal government sets policy mainly through executive orders, and a December 2025 order seeks one national standard and challenges state AI laws it views as too burdensome⁵.

How strict is the EU AI Act?

It sets detailed duties for high-risk systems, such as AI used in hiring, credit and policing². Fines can reach EUR 15 million or 3% of worldwide turnover for many breaches, and more for banned practices⁷.

Related terms

Sources

Back to the glossary