Skip to main content

Knowledge Center

Does UK GDPR Apply to an App Built With AI?

Alex

Alex

CEO

November 6, 2026
7 min read

Short Answer

Yes. UK GDPR applies to personal data, not to the way the software holding it was made. If an app you built with an AI tool stores or handles customers’ or staff names, emails, phone numbers, addresses or anything more sensitive, your business is the controller of that data and has the same duties it would have for an app a developer wrote by hand. The one difference in practice is that with an AI-built app, nobody may yet know whether those duties are being met.

Why People Assume It Might Not

Most owners who build an app with AI start small. A quote calculator, a booking form, a tool that tidies up enquiries. It feels like a personal project rather than a system, so data protection doesn’t come to mind until the app is already collecting details from the public.

There is also a sense that the AI tool must have taken care of the legal side, because the code it produced looks professional. It hasn’t. The tool wrote what it was asked to write. Who answers for the result is the subject of our article on who is responsible for software built with AI, and the short version is that it is the business that put the app in front of people.

What UK GDPR Asks of the App

The section that matters most for software is Article 32, “Security of processing”. It asks the controller to put in place technical and organisational measures appropriate to the risk, taking into account the state of the art, the cost, and the nature and purpose of the processing. It names four kinds of measure:

  • Pseudonymisation and encryption of personal data, where appropriate.
  • The ability to keep systems confidential, intact, available and resilient on an ongoing basis.
  • The ability to restore access to personal data in good time after an incident.
  • A process for regularly testing, assessing and evaluating whether those measures work.

The law doesn’t list specific controls. The ICO’s guide to data security says that “the UK GDPR does not define the security measures that you should have in place”, and points to Cyber Essentials as a sensible starting point for basic technical controls. What you are expected to do scales with what the app holds and the harm a leak would cause.

The last item on that list is the one AI-built apps most often miss. Testing whether your measures work assumes someone knows what the measures are. If the code has never been read, there is nothing to test against.

Where AI-Built Apps Tend to Fall Short

AI coding tools are good at producing working software quickly, and they do what you describe. The gaps sit in the things nobody described. In July 2025, Veracode published a study of AI-generated code that tested over 100 large language models across Java, Python, C# and JavaScript and found that 45% of the code samples introduced a security flaw from the OWASP Top 10, the industry’s standard list of the most common web application risks.

In a small business app, those flaws usually look like this:

  • Records visible to the wrong person. The app checks that someone is logged in but not whether the record they asked for belongs to them, so changing a number in the address shows another customer’s details.
  • Keys in the browser. A database or payment key written into code that runs on the visitor’s device, where anyone can read it.
  • Admin pages left open. A page for managing bookings or exporting contacts that has no login, because the app was only ever tested by its builder.
  • No way back after a failure. No backups, or backups nobody has tried restoring, which runs straight into the “restore access in good time” part of Article 32.

None of these show up when you click through the app yourself. Our explainer on whether vibe coding is bad covers where the approach is fine and where it bites.

The Other Duties That Come With It

Security is the headline, but an app that holds personal data brings a few more obligations with it.

Breach reporting. If personal data leaks and the breach is likely to put people at risk, you have to report it to the ICO within 72 hours of becoming aware of it. That is hard to do well when nobody can say where the app stores its data or who could have seen it.

Suppliers that handle the data. Most AI-built apps sit on a hosting platform, a database service and often an AI model provider that sees what users type. The ICO’s guidance is clear that you remain responsible for data a processor handles for you, so you need to know which services those are and what their terms say.

Apps children might use. If the app is likely to be used by children, the ICO’s Children’s code, formally the Age appropriate design code, sets out fifteen standards for online services. A club booking tool or a tutoring app built over a weekend can fall into this without anyone noticing.

A Worked Example

Take a beauty salon owner who uses an AI tool to build an online booking page. Customers pick a treatment, choose a slot and leave their name, mobile number and a note about allergies. It works on the first try, and bookings start coming in.

Under UK GDPR the salon is now the controller of personal data, and the allergy notes are health information, which the law treats as special category data with stricter conditions. The questions Article 32 raises are practical ones. Is the booking list behind a login? Can a customer see anyone else’s booking by editing the page address? Where is the database, and is it backed up? Which services can read the allergy notes? The owner can’t answer any of them from the screen, because the answers are in the code.

How to Check Yours

Start with what you can do yourself. Write down what personal data the app collects, where it is stored, and every service it passes through. Then work through our guide on how to get a vibe-coded app ready for real users, which lists the checks an owner can run in an afternoon.

If the app holds data that would hurt people if it leaked, it is worth having a developer read the code. Our AI Sign-Off reviews an AI-built app against a written brief, checks its security and data handling against the UK GDPR basics, fixes what fails and signs off the version reviewed. It is a scoped review of a named version, not a certificate of compliance, and it doesn’t replace legal advice on your wider data protection duties. For a larger system that needs an attacker’s view rather than a code review, security testing is the better fit.

Common Mistakes

The most common is treating data protection as something to sort out once the app is popular. The duties apply from the first real customer record. Close behind is assuming that because the AI tool or hosting platform is a large company, security is their problem. They secure their platform; what your app does with the data on it is yours.

The last is fixing one problem by prompting the tool again and assuming the rest is fine. Each change is new, unreviewed code. More answers on related topics are in the Knowledge Center.

Disclaimer: The information provided in this article is for general guidance only and does not override or replace any terms in your contract. While we aim to offer helpful insights through our Knowledge Center, the accuracy of content in this section is not guaranteed.

Written by

Alex

CEO

I’m a software developer and CEO of Digital Royalty, helping growing teams scale their SaaS platforms without losing quality, visibility, or control. I focus on building structured, maintainable systems with clear processes, reporting, and accountability. With over a decade of experience across agency and in-house roles, I specialise in delivering long-term, scalable solutions that support complex, evolving products.

Portrait of Alexander De Sousa, founder of Digital Royalty
Founder-led
“I’ve put everything I know into how this company works — the standards, the method, the care on every project. It runs through the whole team, and I hold us all to it.”

Alexander De Sousa · Founder LinkedIn

Featured on BBC Radio Solent

Get started

Tell us what you need

A few quick questions, then a straight answer from a real person — usually within a few hours.

What happens next

  1. 1 We reply A real person reads your answers and gets back to you, usually within a few hours.
  2. 2 Questions or a short call If anything isn't clear, we ask by email or book a short call. Before any call we audit your website and make the quick fixes it finds, free.
  3. 3 Your fixed quote A clear, fixed price in writing, within 48 hours of any call. There is no obligation to go ahead.

Tell us what you're working on

Whether it's a new site, a platform, or a process that shouldn't be manual any more — we'll tell you honestly if we can help.