The UK approach to AI regulation
White paper "A pro-innovation approach to AI regulation" (2023) and Government response (2024) · Official source (GOV.UK)
The United Kingdom has no AI law. It chose an approach based on five cross-sector principles that existing regulators (privacy, competition, finance, communications, health) apply in their own sectors. General laws remain fully applicable: data protection, equality, consumer protection and online safety.
⚠️ The Government has repeatedly announced a law on the most powerful AI models, but as of November 2026 it has not been introduced. The Regulating for Growth Bill provides for regulatory sandboxes, including for AI.
Who it applies to
- Anyone developing or using AI systems in the UK, through existing sector laws
- Sector regulators, tasked with applying the five principles
- Online services with generative AI features, which fall under the Online Safety Act
Key obligations
The five principles
Safety and robustness; transparency and explainability; fairness; accountability and governance; contestability and redress.
Data protection
The UK GDPR governs the use of personal data to train and use AI, and automated decisions.
Non-discrimination
The Equality Act 2010 also applies to decisions taken or supported by AI systems.
Online safety
The Online Safety Act imposes duties on services, including chatbots, that allow content to be generated or shared.
Regulated sectors
Finance (FCA), health (MHRA), competition (CMA) and communications (Ofcom) apply their own AI guidance.
Penalties
No AI-specific penalties
Sector law penalties apply: for example up to £17.5M or 4% of turnover under the UK GDPR and up to £18M or 10% of turnover under the Online Safety Act.
Competent authorities
Sector regulators (ICO, CMA, FCA, Ofcom, MHRA) coordinated by the Digital Regulation Cooperation Forum; Department for Science, Innovation and Technology (DSIT); AI Security Institute for the most advanced models.