Legal duty
Reporting an AI data breach under UK GDPR
An AI data breach is reportable when it is a personal data breach likely to result in a risk to people. The deadline is 72 hours from becoming aware, where feasible.
Last reviewed against legislation.gov.uk, National Cyber Security Centre and OWASP GenAI Security Project. Sources
The rule
What Article 33 says
UK GDPR Article 33 says the controller must notify the regulator of a personal data breach promptly and, where feasible, not later than 72 hours after becoming aware of it [1]. Notification is not required if the breach is unlikely to result in a risk to the rights and freedoms of individuals [1]. If notification is late, the reasons for the delay must be given [1].
The article applies to personal data wherever it was held. It makes no exception for AI tools, so a prompt log full of customer details leaked by a supplier is treated like any other leak of that data.
Applying it
Three questions
Was personal data involved
Prompts, chat history, uploaded files and outputs may all hold names, contact details or case details. Establish what the exposed data contained before deciding.
Is a risk likely
The test is whether the breach is unlikely to result in a risk to people. Record your reasoning, including what the data was, who could have seen it and what you know about its use.
When did you become aware
The clock runs from the point you become aware of the breach, not from the end of your investigation [1]. Note the time and who knew.
Record it
Whether or not you report
The controller must document any personal data breach, including the facts, its effects and the remedial action taken [1]. Keep a simple log with the date you became aware, the data involved, the decision and who made it.
Where the processor is an AI supplier
If an AI supplier holds the data on your behalf, check the contract for how and when they must tell you. You cannot act inside 72 hours on a breach you do not hear about, so ask the question early.
Next steps
In order
-
1
Contain the exposure using the containment steps. -
2
Establish what personal data was involved. -
3
Decide whether it is reportable and note the reasons. -
4
If reportable, report through the ICO's own channel, and check its current guidance on whether and how to tell affected people. -
5
Update your record and review what failed.
Keep reading
Read next
Examples
Three public AI data exposure cases, a library bug, staff uploads and an open database, with what each shows UK organisations about cause and response.
Read the guide
Prevention
Practical controls to stop data leaking through AI tools and pipelines, drawn from NCSC guidance, OWASP and IBM research, in order of priority.
Read the guide
Containment
A practical first-response plan for a suspected AI data breach: stop the exposure, scope the data, preserve evidence, decide on reporting and review.
Read the guideHow CyPro can help
Need a hand working through this?
Finding out whether an incident reaches you is one job. Checking your own systems, tightening access and briefing the board are several more. CyPro's responders and consultants can take any of them on.
Incident response
If you may already be compromised, our responders contain it, preserve the evidence, find out what was touched and get you running again safely.
More on cypro.co.uk
24/7 security monitoring
Our SOC analysts watch your systems day and night for the signs of an attack in progress: unexpected processes, new outbound connections, logins that do not fit.
More on cypro.co.uk
Fractional CISO for the board
A senior security lead who advises the board on what to fix first, what to accept and what to tell customers, and owns the decision record.
More on cypro.co.uk
Penetration testing
A focused test of your systems and the controls around them, so you know they hold before an attacker finds out for you.
More on cypro.co.uk
Hardening and segmentation
Our consultants put the controls in place that limit the damage when something is missed: firewall rules, network isolation and tighter admin access.
More on cypro.co.uk
Vulnerability scanning
Find every internet-facing system you run, including the forgotten ones, and see which are still on a vulnerable build.
See the scanning service
Talk to a specialist
Worried this affects you?
If data has leaked through an AI tool or an exposed AI database, the first hours decide how much is lost and what you must report. An incident response team helps you contain it, work out what was exposed and meet your notification duties.