Response
How to contain an AI data breach
Stop the exposure first, then find out what data was involved, because that decides your reporting duty. The NCSC advises having a response plan for serious AI-related incidents before you need one.
Last reviewed against legislation.gov.uk, National Cyber Security Centre and OWASP GenAI Security Project. Sources
Before you start
Awareness starts the clock
UK GDPR counts 72 hours from when you become aware of a reportable personal data breach [1]. Write down the time you were first told, by whom, and what they said. Everything else follows from that note.
The steps
In order
-
1
Stop the exposure. Close the open port, revoke the access, disable the integration or suspend the tool. -
2
Rotate secrets. API keys and tokens that appeared in logs or prompts should be treated as exposed. Wiz found API keys in the exposed DeepSeek logs [6]. -
3
Preserve evidence. Keep logs, configuration and screenshots before you change anything further. -
4
Scope the data. Establish what categories of data were involved and whose they are. -
5
Contact the supplier, if one is involved, and ask for their account of what happened. -
6
Decide on reporting, and record the reasons [1]. -
7
Brief leadership and staff with facts only.
Judgement calls
Where teams get stuck
Is it really a personal data breach
If prompts or outputs held names, contact or account details, yes, they are personal data. The ChatGPT case involved names, email and payment addresses, and card details in part [7].
Can we ask for the data back
Samsung's memo said data entered into AI platforms is hard to retrieve or delete [8]. Ask the supplier what deletion they can confirm in writing, and do not assume.
Who speaks to customers
After
Learn from it
Add the cause to your risk list and update the prevention controls. Review the reporting record is complete.
Keep reading
Read next
Suppliers
Questions to ask AI suppliers about data storage, training use, access and incident notification, based on NCSC guidance for UK organisations.
Read the guide
Data protection
What the ICO's AI and data protection guidance says about impact assessments, transparency and lawfulness, and how it links to breach duties.
Read the guide
AI use policy
How to write a short AI use policy that stops sensitive data going into AI tools, using OWASP, NCSC and IBM findings as the basis.
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.