AI regulation around the world: EU, US, UK and India compared
The EU has a single, risk-based AI law, the AI Act, with transparency rules applying since 2 August 2026 and high-risk rules from December 2027. The US has no comprehensive federal AI law; states such as California, Colorado and Texas have their own, while the federal government pushes to preempt them. The UK regulates AI through existing regulators and laws, with no AI bill before Parliament. India relies on IT rules, including mandatory labelling of synthetic media since February 2026, plus data protection law and voluntary guidelines.
How do the four approaches compare?
AI regulation has split into distinct models. The EU wrote a single horizontal law. The US relies on existing federal agencies and a growing set of state laws. The UK chose to empower existing regulators rather than pass an AI act. India regulates through its information technology rules and data protection law. For businesses, the key point is that rules usually follow your users: an AI product used in the EU must meet EU rules even if the company is based elsewhere.
| Region | Main approach | Status for businesses |
|---|---|---|
| EU | Single risk-based law (AI Act) | Prohibitions, AI literacy, GPAI and transparency rules apply; high-risk rules from Dec 2027 and Aug 2028 |
| US | State laws plus existing federal powers | No comprehensive federal statute; state laws in force; federal preemption contested |
| UK | Existing regulators and laws | No AI bill before Parliament; data protection and sector rules apply |
| India | IT Rules, data protection, guidelines | Synthetic media labelling mandatory since 20 Feb 2026 |
What does the EU AI Act require?
The AI Act classifies AI by risk. Some practices are banned; high-risk systems such as those used in hiring, credit, education and medical devices face strict obligations; and transparency duties apply to chatbots, synthetic content and deepfakes. The Digital Omnibus (Regulation 2026/1744), in force since 27 July 2026, moved standalone high-risk obligations to 2 December 2027 and product-embedded ones to 2 August 2028, but left the Article 50 transparency rules applying from 2 August 2026. Our EU AI Act guide covers the detail.
How is AI regulated in the US?
There is no comprehensive federal AI statute. Federal agencies use existing powers on consumer protection, discrimination and safety. States have moved faster: California’s SB 53 sets transparency duties for frontier model developers, Texas has its TRAIGA law, and Colorado replaced its original AI Act with a narrower law effective January 2027. New York City requires bias audits for automated hiring tools. A December 2025 executive order and a March 2026 White House framework push for federal preemption of state laws, but the framework is non-binding and state laws remain in force unless Congress acts or courts intervene.
How is AI regulated in the UK?
The UK has no AI act, and no AI bill is before Parliament. Instead, existing regulators such as the ICO, FCA, CMA and MHRA apply their rules to AI in their sectors. The Data (Use and Access) Act 2025 updated rules on automated decision-making from February 2026, and the ICO is preparing a statutory code of practice on AI and automated decision-making. The Advertising Standards Authority applies its existing codes to AI-generated ads. A Regulating for Growth Bill announced in 2026 focuses on regulatory sandboxes rather than new AI obligations.
How is AI regulated in India?
India has no dedicated AI act. Amendments to the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, effective 20 February 2026, require realistic AI-generated audio, images and video to be clearly labelled and carry provenance metadata, with fast takedown duties for platforms. The Digital Personal Data Protection Act governs personal data used by AI systems, and the government’s AI Governance Guidelines from November 2025 set voluntary principles. The advertising self-regulator ASCI has also issued draft guidance on labelling AI-generated ads.
How can one business comply with all four?
- Inventory your AI: List every AI system you build or use, where users are and what it decides.
- Classify by risk: Use the EU categories as a baseline; they are the most detailed.
- Meet transparency everywhere: Disclose chatbots, label realistic synthetic media and mark generated content.
- Add local rules: State hiring rules, India labelling, UK sector guidance.
- Keep humans in control: Review points for decisions about people.
- Monitor changes: Assign an owner and review every quarter.
Our guides to AI disclosure rules for marketers, AI usage policies and human-in-the-loop AI help put this into practice.
What about the UAE and Gulf?
Many of our clients are based in the UAE, which has a national AI strategy, a federal personal data protection law and free-zone regimes such as DIFC with their own data rules, including provisions on autonomous systems. Businesses serving EU, UK, US or Indian users from the Gulf still need to meet those markets’ rules.
How UIUX Studio helps
We help teams design compliant AI products, from disclosure patterns to review workflows. Talk to our Technology Advisory or Agentic AI & Automation teams. This article is general information, not legal advice.
Frequently asked questions
Which country has the strictest AI law?
The EU, with the AI Act’s detailed risk-based rules.
Is there a US federal AI law?
No comprehensive one; states and existing agencies regulate AI.
Does the UK have an AI act?
No; existing regulators apply their rules to AI.
Does India require AI labels?
Yes, for realistic AI-generated audio, images and video since February 2026.
Do EU rules apply to non-EU companies?
Yes, when their AI is used in the EU.