AI regulation in the UK: what applies in 2026
Written by Alex Vakhitov, founder of Comonad Limited. Last reviewed: .
On this page
Is there a UK AI Act?
No. In its March 2023 white paper, "A pro-innovation approach to AI regulation", the government chose to regulate AI through existing regulators, with no new cross-sector law, and its 6 February 2024 response (opens in new tab) confirmed that approach.
The five principles regulators apply
Safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; contestability and redress. They are not law in themselves. Each regulator interprets them within its own remit, and DSIT has published guidance for regulators (opens in new tab) on doing so.
Data protection: UK GDPR and the ICO
If your AI uses personal data, UK GDPR applies, and the Information Commissioner's Office is the regulator. The ICO's Guidance on AI and data protection (opens in new tab) covers lawful basis, fairness, transparency, accuracy, data minimisation and security in AI systems. Solely automated decisions with legal or similarly significant effects carry extra safeguards, and the Data (Use and Access) Act 2025 changed those rules from 5 February 2026. A data protection impact assessment is required where the processing is likely to result in a high risk to people.
Sector regulators
These are the regulators whose existing powers most often touch AI. Read the publications that match your sector.
| Regulator | Where it touches AI |
|---|---|
| ICO | Personal data in AI systems; automated decision-making; guidance on AI and data protection |
| FCA | AI used by financial services firms, under existing rules such as the Consumer Duty and the Senior Managers and Certification Regime |
| CMA | Competition and consumer protection in AI markets, including its work on AI foundation models |
| Ofcom | Online services and broadcasting. The Online Safety Act applies to in-scope services, including some that use generative AI, and to AI-generated content shared between users. |
| Ofgem | AI in the energy sector; its "Ethical AI use in the energy sector" guidance, updated May 2026 |
| MHRA | AI as a medical device |
| DRCF | The Digital Regulation Cooperation Forum (opens in new tab), through which the ICO, CMA, FCA and Ofcom coordinate on digital regulation, including AI |
Government guidance worth knowing
DSIT published its AI Risk Management Toolkit (opens in new tab) on 8 September 2026, for anyone designing, procuring or delivering an AI product. The NIST AI Risk Management Framework and ISO/IEC 42001 are voluntary. Teams use them to structure controls.
How the EU AI Act reaches UK firms
Separately from UK law, the EU AI Act applies to UK organisations whose AI is placed on the EU market, put into service there, or whose output is used in the EU. Its high-risk duties start on 2 December 2027 and 2 August 2028. See Does the EU AI Act apply to UK companies?.
What a UK organisation should do now
- Keep an inventory of the AI you build, buy and use.
- For each system that uses personal data, check the lawful basis and whether a DPIA is needed.
- Identify the regulators whose rules apply to your sector and read their AI publications.
- Mark any system whose AI or output reaches the EU, and check it against the EU AI Act.
- Name an owner for each system and a review cycle.
How Comonad helps
Our AI governance consultancy turns these duties into one control set, with owners, tests and records, mapped to UK regulatory guidance and the EU AI Act where it applies.
Discuss your AI governanceSources
- AI regulation: a pro-innovation approach (29 March 2023) (opens in new tab)
- Government response, 6 February 2024 (opens in new tab)
- Implementing the UK's AI regulatory principles (DSIT) (opens in new tab)
- Data (Use and Access) Act 2025, section 80 (opens in new tab)
- Data (Use and Access) Act 2025 (Commencement No. 6) Regulations 2026 (opens in new tab)
- ICO guidance on AI and data protection (opens in new tab)
- FCA: our approach to AI (opens in new tab)
- CMA: AI foundation models initial report (opens in new tab)
- Ofcom: AI chatbots and online regulation (opens in new tab)
- Ofgem: ethical AI use in the energy sector (opens in new tab)
- Medicines and Healthcare products Regulatory Agency (opens in new tab)
- Digital Regulation Cooperation Forum: about us (opens in new tab)
- DSIT AI Risk Management Toolkit (opens in new tab)
Frequently asked questions
Is there an AI regulator in the UK?
No single one. The ICO, FCA, CMA, Ofcom, Ofgem and MHRA each apply existing law within their own remit. The ICO, CMA, FCA and Ofcom coordinate through the Digital Regulation Cooperation Forum.
Does UK GDPR apply to AI?
Yes, wherever personal data is processed. That includes training data, prompts, retrieved context and outputs that identify a person. The ICO's guidance on AI and data protection is the starting point. A data protection impact assessment is required where the processing is likely to result in a high risk to people.
Is the UK going to copy the EU AI Act?
There is no UK equivalent of the EU AI Act. The government's 6 February 2024 response kept regulation with existing regulators and did not adopt a new cross-sector law.