← Journal

AI and UK data protection: what regulated firms must check

Two professionals reviewing a data protection checklist beside a laptop

Professional firms hold some of the most sensitive personal data there is: income, health, family circumstances, disputes and wealth. Before any of it goes into an AI tool, whether a chatbot, a transcription service or an automated intake system, the firm needs to be confident it is meeting UK GDPR.

The ICO has published detailed guidance on artificial intelligence and data protection, including guidance on explaining decisions made with AI and an AI and data protection risk toolkit. This post turns the main points into a practical checklist for accountants, financial advisers and law firms. It is not legal advice, and your data protection lead should own the final assessment.

1. Know what data the tool will see

Start by listing exactly what personal data will be processed. An intake chatbot may collect names, contact details and a description of the enquiry. A meeting transcription tool may capture everything said in a client review, including health or family information, which can be special category data with stricter conditions.

Collect only what you need. If the tool can do its job without date of birth or account numbers, do not give them to it.

2. Identify the lawful basis

Every processing activity needs a lawful basis under UK GDPR. For enquiry handling, that is often legitimate interests or steps taken at the person's request before entering a contract. For special category data, you also need an additional condition. Record which basis you rely on and why.

3. Decide whether you need a DPIA

UK GDPR requires a data protection impact assessment where processing is likely to result in a high risk to individuals. New technologies, large volumes of sensitive data and automated decision-making are all factors. Many uses of AI with client data will meet that threshold, and the ICO's AI and data protection risk toolkit is designed to help you work through the risks. If in doubt, do one. It is also good evidence of a considered decision.

4. Check the supplier

Most firms use AI through a third-party supplier, which usually makes the supplier a processor acting on your instructions. You need a written contract that meets UK GDPR requirements. Ask:

  • Where is the data stored and processed?
  • Is client data used to train the supplier's models? Can that be switched off?
  • How long is data retained, and can it be deleted on request?
  • What security measures and certifications does the supplier have?
  • Which sub-processors does it use?

Free consumer versions of AI tools often have very different terms from business versions. Staff should not paste client data into a personal account.

5. Check international transfers

If the supplier processes data outside the UK, you need a valid transfer mechanism, such as a UK adequacy decision or appropriate safeguards like the International Data Transfer Agreement. Many AI suppliers are based in the US, so this question comes up often.

6. Watch for automated decisions

UK GDPR gives individuals rights relating to decisions based solely on automated processing that have legal or similarly significant effects. For most professional firms, the safest design is to keep a qualified person in every decision that matters: AI can gather information, summarise and route enquiries, but a person decides whether to take on a client, what to recommend and how to respond to a complaint.

7. Be transparent

Update your privacy notice to explain how you use AI, in plain language. If clients interact with an automated system, such as a website chat or an AI receptionist, tell them it is automated and how to reach a person. The ICO's guidance on explaining decisions made with AI covers what good explanations look like.

8. Keep accuracy in mind

AI tools can produce confident but wrong outputs. If an AI summary becomes part of a client record, someone should check it. Inaccurate personal data in a file is a data protection problem as well as a professional one.

9. Respect your professional rules too

Data protection is only one layer. Solicitors have duties of confidentiality, accountants have professional body rules on confidentiality, and FCA-regulated firms have record-keeping and systems and controls requirements. An AI tool that is acceptable under UK GDPR still needs to fit those obligations.

Common questions

Can staff use free AI chatbots with client information?

Generally they should not. Free consumer versions often have terms that allow data to be retained or used to improve the service, and there is no processor contract in place. Firms should provide approved business tools and a clear policy on what may and may not be entered.

Is a website chatbot covered by UK GDPR?

Yes, if it collects personal data, which most do when someone types their name, email or the details of their enquiry. It needs a lawful basis, a mention in your privacy notice, a compliant supplier contract and sensible retention settings.

Do we always need a DPIA for AI?

Not always, but often. A DPIA is required where processing is likely to be high risk, and AI used with sensitive client data frequently meets that test. Where the decision is borderline, completing one is a sensible way to show you considered the risks.

A short checklist

  1. List the personal data the tool will process and minimise it.
  2. Record the lawful basis and any special category condition.
  3. Complete a DPIA where the processing is likely to be high risk.
  4. Put a compliant processor contract in place and confirm training and retention settings.
  5. Confirm a valid mechanism for any international transfer.
  6. Keep a qualified person in decisions with significant effects.
  7. Update the privacy notice and label automated interactions.
  8. Check AI outputs before they enter client records.
  9. Confirm the tool fits your professional and regulatory obligations.

We build intake and follow-up systems for regulated firms with these questions answered before anything goes live. See AI for accountants, AI for lawyers and AI for financial advisors, or our AI automation agency page for how we approach projects.