Case studies
AI data breach examples and what they show
Three well-documented cases show three different causes. None is a model being hacked; each is a familiar failure applied to AI.
Last reviewed against legislation.gov.uk, National Cyber Security Centre and OWASP GenAI Security Project. Sources
Why look at cases
Causes differ
A UK leader asking "could this happen to us" gets a better answer from specific cases than from a general warning. The three below are the best documented in public.
The cases
Side by side
| Case | Date | Cause | Data involved | Source |
|---|---|---|---|---|
| ChatGPT | 20 March 2023 | Bug in the open-source redis-py library | Other users' chat queries and some subscribers' payment details | [7] |
| Samsung | April 2023, memo reported 2 May 2023 | Staff uploaded internal source code | Source code | [8] |
| DeepSeek | Reported 29 January 2025 | Database exposed without authentication | Over a million log entries, chat history, API keys | [6] |
ChatGPT
A supplier bug
OpenAI said the bug showed other users' chat queries and personal information for about 1.2% of ChatGPT Plus subscribers active in a nine-hour window [7]. The exposed details included name, email address, payment address, the last four digits of a card and its expiry date, and OpenAI said full card numbers were not exposed [7]. OpenAI said it was contacting everyone whose payment information was exposed [7].
The lesson for customers: your exposure can come from a supplier's dependency, so you need to know how a supplier will tell you.
Samsung
A staff decision
Engineers uploaded internal code to get help with errors [8]. The company's response was a restriction on generative AI tools and work towards internal tools and ways to block uploads of sensitive information [8]. The lesson: if there is no approved tool and no rule, staff will pick one. See shadow AI data leaks.
DeepSeek
An open system
The exposure was a missing password on a database, not anything specific to AI [6]. Wiz said DeepSeek secured it promptly after disclosure [6]. The lesson: AI products inherit every ordinary infrastructure risk. See exposed AI databases.
What these cases do not show
The sources do not say how many people were harmed in any of the three, so this site makes no claim about harm.
Keep reading
Read next
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 guide
Suppliers
Questions to ask AI suppliers about data storage, training use, access and incident notification, based on NCSC guidance for UK organisations.
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.