AEVOMIND INSIGHTS — AI & THE LAW

The UK rules on AI for businesses: data protection, automated decisions and transparency

19 October 2026 · Updated 15 November 2026 · 7 min read · By the AevoMind team

A clear glass padlock with a glowing gold keyhole, threaded with fine streams of green light, above a dark glossy floor

There is no single UK AI law. AI is governed by the laws that already apply, data protection above all, with regulators applying five shared principles. For most businesses that means an impact assessment before you start, clear information for the people affected, and specific safeguards if AI makes significant decisions about people on its own.

The short answer

There is no single UK AI law. AI is governed by the laws that already apply to what you are doing — data protection above all, plus equality, consumer and sector rules — with regulators applying five shared principles. For most businesses, the obligations that matter are a data protection impact assessment before you start, clear information for the people affected, and, if the AI makes significant decisions about people on its own, specific safeguards that changed in 2026.

This article explains each in plain terms. It is not legal advice; take your own on your circumstances.

The UK approach: five principles, existing regulators

The government’s response on AI regulation (February 2024) set out five cross-sector principles for existing regulators to interpret and apply within their remits:

  1. Safety, security and robustness.
  2. Appropriate transparency and explainability.
  3. Fairness.
  4. Accountability and governance.
  5. Contestability and redress.

So the questions a regulator will ask about your AI are the questions it already asks about your business, applied to a new tool.

A frosted-glass panel on a black base with soft gold shapes behind it, and a brass magnifying glass resting in front
Look closely before you deploy: the impact assessment comes first.

Data protection: the impact assessment

ICO guidance is direct: “In the vast majority of cases, the use of AI will involve a type of processing likely to result in a high risk to individuals’ rights and freedoms, and will therefore trigger the legal requirement for you to undertake a DPIA.” Where you decide a particular use is not high risk, you still need to document how you reached that view.

A good assessment answers: what personal data the AI uses and why, where it goes, who can see it, how long it is kept, what could go wrong for the people involved, and what you do to prevent it.

Three glass gateways edged in gold light standing over glowing green paths on a dark floor
Checks before the decision, not apologies after it.

Automated decisions: what changed in 2026

The Data (Use and Access) Act 2025 rewrote the UK GDPR rules on decisions made by machine. According to the ICO, organisations can now use a much wider range of lawful bases, including legitimate interests, for significant decisions made solely by automated means, provided safeguards are in place; the new recognised legitimate interests basis cannot be used for such decisions (Article 22B), and special category data stays more tightly restricted. The ICO says all of the Act’s data protection provisions were in force by 19 June 2026, and it is working on a statutory code of practice on AI and automated decision-making.

Where a significant decision is based solely on automated processing, Article 22C requires safeguards that:

The simplest compliant design is often to keep a person in the decision: the AI prepares, a person decides. Keep a record of both — see audit trails.

Three blank frosted-glass cards fanned out on a black surface beside a brushed gold pen
Write it down: what the AI may do, what it may not, and who is accountable.

The AI your staff already use

Rules apply to the AI your business has not chosen as well as the AI it has. The Office for National Statistics found 55% of working people in Great Britain using AI for work or education in May to June 2026, while 34.9% of UK businesses with ten or more employees reported using it. The surveys differ, but the gap is a reason to write an AI use policy.

More working people use AI than businesses report using itEmployed and self-employed people using AI for work or education (Great Britain, May to June 2026) and businesses with 10 or more employees using AI (UK, June 2026)
Working people using AI for work or education55.0%
Businesses using AI34.9%

Source: Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026 (July 2026), Business Insights and Conditions Survey. Open Government Licence v3.0. The two figures come from different surveys and are not directly comparable; the employee figure may include some use outside formal business activity.

Show the data
Share
Working people using AI for work or education55.0%
Businesses using AI34.9%

And look at what the AI is used for. Customer-facing uses touch more personal data and more people than internal ones, so they deserve the closest look in an impact assessment.

What small businesses use AI forAmong UK businesses with 10 to 49 employees using AI, share using it for each purpose, June 2026
Improve business operations56.3%
Provide or personalise products or services for customers23.5%
Develop a new product or service16.1%
Explore a new market11.3%

Source: Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026 (July 2026), Business Insights and Conditions Survey. Open Government Licence v3.0. Businesses may report more than one purpose.

Show the data
Use AI for this
Improve business operations56.3%
Provide or personalise products or services for customers23.5%
Develop a new product or service16.1%
Explore a new market11.3%

A practical checklist

  1. List every AI tool in use, including the ones staff adopted themselves.
  2. Write a short AI use policy: what may be put into which tools, and what never may.
  3. Do the impact assessment for each use involving personal data, or record why it is not needed.
  4. Tell people plainly where AI is involved and how to reach a person.
  5. Keep a person in significant decisions, or put the Article 22C safeguards in place.
  6. Check for fairness where AI touches hiring, pricing or eligibility; equality law still applies.
  7. Keep records of what the AI did and who checked it.

If you are at the start, our guide to implementing AI in a UK business puts these steps in order; for public bodies, see AI in UK government.

A glowing gold outline of an open hand resting beside a clear glass control dial on a dark surface, with fine threads of green light
A person stays in control of the decisions that matter.

Frequently asked questions

Is there a UK AI law businesses must follow?

The UK has not passed a single, general AI law of the kind the EU adopted. Instead the government asked existing regulators to apply five cross-sector principles within their own areas: safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress. Existing law, especially data protection, still applies to AI in full.

Does using AI need a data protection impact assessment?

In most cases, yes. ICO guidance says that in the vast majority of cases the use of AI will involve processing likely to result in a high risk to individuals' rights and freedoms, which triggers the legal requirement for a data protection impact assessment. If you conclude a particular use is not high risk, the ICO says you still need to document how you reached that view.

What changed for automated decision-making in 2026?

The Data (Use and Access) Act 2025 rewrote the UK GDPR rules. Organisations can now rely on a much wider range of lawful bases, including legitimate interests, for significant decisions made solely by automated means, as long as safeguards are in place; the new recognised legitimate interests basis cannot be used for such decisions, and special category data stays more tightly restricted. The ICO says all of the Act's data protection provisions were in force by 19 June 2026.

What safeguards must an automated decision have?

Under Article 22C of the UK GDPR, where a significant decision is based solely on automated processing, the safeguards must include giving the person information about the decision, enabling them to make representations, enabling them to obtain human intervention, and enabling them to contest the decision. A person who can review and change the outcome is the core of it.

Should we tell customers when AI is involved?

In our view, yes, plainly and early. Transparency is one of the five UK principles and a core data protection duty, and customers who find out later tend to feel misled. Say what the AI does, what it does not decide, and how to reach a person. Take your own advice on the exact wording for your circumstances.