Using AI and UK GDPR: what businesses need to know

How UK data protection law applies when your business uses AI tools, chatbots and automation, and practical steps to stay compliant.

By the UK Technology Hub teamPublished 4 minute read

The short answer

If your AI tools process personal data, UK GDPR applies. You need a lawful basis, must tell people how their data is used, collect only what you need, keep it secure, check what your AI suppliers do with it, and be careful with automated decisions that significantly affect people. For higher-risk uses, carry out a data protection impact assessment. The ICO publishes detailed guidance on AI.

AI tools are only as safe as the way they're used. When AI handles customer enquiries, staff information or documents containing personal details, data protection law applies. This guide explains the main points in plain English. It's general information, not legal advice.

Does UK GDPR apply to your AI use?

UK GDPR and the Data Protection Act 2018 apply whenever you process personal data, meaning information about an identifiable living person. That includes names, emails, phone numbers, chat transcripts, uploaded documents, voice recordings and even opinions about someone. If your AI tool touches any of these, the rules apply.

The Information Commissioner's Office (ICO) is the UK regulator and publishes guidance on AI and data protection. The Data (Use and Access) Act 2025 updated parts of UK data protection law, so check the ICO's current guidance for the latest position.

The key principles applied to AI

Lawful basis. You need a valid reason to process personal data, such as legitimate interests or performing a contract. Using customer data to train or improve AI may need a different basis from using it to answer their question.

Transparency. Tell people when they're interacting with AI and explain how their data is used, in your privacy policy and at the point of collection. A short line in a chatbot such as "You're chatting with our AI assistant. See our privacy policy." is a good start.

Purpose limitation. Use data only for the purposes you've told people about.

Data minimisation. Collect only what's necessary. A chatbot answering product questions doesn't need someone's date of birth.

Accuracy. AI can produce incorrect statements. Where outputs about individuals matter, have a process to check and correct them.

Storage limitation. Decide how long chat logs and processed documents are kept, and delete them afterwards.

Security. Protect data with access controls, encryption and secure suppliers.

Choosing AI suppliers

Most businesses use AI through suppliers such as AI model providers, chatbot platforms or automation tools. Before using one, check:

  • Whether they use your data to train their models, and whether you can opt out. Business and API plans of major providers commonly exclude customer data from training by default, but confirm this.
  • Where data is stored and processed, and whether international transfers have appropriate safeguards.
  • Their data processing agreement, which UK GDPR requires when a supplier processes personal data on your behalf.
  • Retention settings for prompts and outputs.
  • Security certifications and incident response.

Automated decisions

UK GDPR contains specific rules on decisions made solely by automated means that have legal or similarly significant effects on people, such as rejecting a job application, a loan or a tenancy. These rules were updated by the Data (Use and Access) Act 2025, but safeguards remain, including telling people about such decisions and allowing them to contest them and obtain human involvement. The simplest approach for most small businesses is to keep a person responsible for significant decisions, with AI assisting.

Data protection impact assessments

A DPIA is required for processing likely to result in high risk to individuals. The ICO considers many uses of innovative technology like AI likely to need one, especially involving large-scale processing, sensitive data, monitoring or significant decisions. A DPIA helps you identify risks and decide how to reduce them.

Staff using AI tools

One of the biggest risks is staff pasting customer or confidential information into consumer AI tools. Create a simple AI policy:

  1. Which AI tools are approved, ideally business plans with appropriate data protections.
  2. What information must never be entered, such as special category data or client confidential material.
  3. That outputs must be checked before use.
  4. Who to ask if unsure.

Chatbots and customer conversations

For customer-facing chatbots, our guide comparing ChatGPT and custom chatbots explains why control matters. Key steps:

  • Tell users they're talking to AI.
  • Link to your privacy policy.
  • Avoid collecting sensitive data unless essential.
  • Set retention periods for transcripts.
  • Provide a route to a human.

If you serve customers in the EU, the EU AI Act may also apply, including transparency duties for chatbots.

A practical checklist

  • List every AI tool in use and what data it touches.
  • Confirm a lawful basis for each use.
  • Update your privacy policy.
  • Review supplier terms and sign data processing agreements.
  • Carry out a DPIA for higher-risk uses.
  • Set retention periods.
  • Train staff and publish an AI policy.
  • Review regularly as tools and guidance change.

How we build with privacy in mind

Our AI solutions and AI chatbot development services design privacy in from the start: choosing suppliers with suitable data terms, minimising data collected, setting retention and adding clear notices. When we add AI features to web apps, we follow the same approach. For ideas on where to begin, see our guide to practical AI uses for small businesses.

Frequently asked questions

Can I put customer data into ChatGPT?

Be very cautious. Check the plan's data terms, whether data is used for training, and whether your privacy notice covers it. Business plans offer stronger protections than personal accounts.

Do I need to tell customers I use AI?

If AI processes their personal data or they interact with it directly, you should be transparent. It also builds trust.

Do I need a DPIA for a chatbot?

It depends on the risk. A chatbot answering general questions may be low risk; one handling sensitive data or making significant decisions likely needs one.

This guide is general information, not legal, tax or financial advice, and details such as prices, features and rules can change. Product and company names mentioned are trademarks of their respective owners and are used only to identify those products; UK Technology Hub is not affiliated with or endorsed by them.

Related guides

All AI & automation guides