Check your exposure
Is your organisation exposed to AI-assisted attacks?
Most organisations are exposed in at least one of three places: people who act on requests, software reachable from the internet, and AI tools they have started to use. Check each in turn.
Last reviewed against NCSC (UK), NCSC New Zealand and UK Government (DSIT and Home Office). Sources
Your people
Who can be impersonated
Ask who in your organisation can move money, change bank details, reset access or share data on the strength of a call, video meeting or message. These are the roles an impersonator targets. In the Arup case the employee first suspected a phishing email, then set doubts aside because the people on the video call looked and sounded like colleagues [7].
Check these as well:
- Whether payment changes can be approved from a single request
- Whether the service desk resets credentials after a phone call alone
- Whether senior staff have public video and audio that could be copied
- Whether staff know how to raise a doubt without embarrassment
Your software
What an attacker can reach
NCSC guidance says the first step is identifying what systems you have and which are internet accessible [4]. It names admin login panels, legacy systems and operational technology as needing extra attention [4]. The reason is speed: the NCSC says the gap between disclosure and exploitation has shrunk to days [1].
Your AI tools
What they can read and do
If anyone in your organisation uses an LLM that reads email, documents or web pages and can also use tools or send data, it is exposed to prompt injection. The NCSC's example is an LLM handling email from unknown external senders, which should not have access to privileged tools [2]. In the 2025/2026 survey, about a third of businesses were using, adopting or considering AI, and only about a quarter of those had security practices for its risks [5].
Quick test
For each AI tool, ask what the worst thing is that a hostile message could make it do. If the answer involves sensitive data or actions you could not undo, restrict it first [2].
Am I affected?
Answer a few questions
Check if you are affected
Answer three questions about your people, your internet-facing systems and your AI tools to see where to look first. Nothing you answer leaves this page.
Can anyone approve a payment, a bank detail change or an access request on the strength of a call, video meeting or message alone?
What is affected
The software and products in scope
Staff who approve payments, reset access or act on video and voice requests[7]
An Arup employee in Hong Kong joined a video call with impersonated senior officers and sent money to the criminals' accounts.
Organisations with internet-facing or unpatched software[4]
AI is shortening the gap between vulnerability discovery and exploitation, so defenders have less time to patch.
Organisations using or adopting AI tools[5]
About a quarter of those in the UK survey had security practices for AI-related risks.
Systems that let an LLM read untrusted content and use tools[2]
Prompt injection may never be fully mitigated, so privileges and impact have to be limited by design.
Critical national infrastructure that integrates AI[1]
The NCSC says AI integration creates a larger attack surface.
Open-source code and software supply chains[4]
The NCSC guidance says open-source code is likely to be heavily scanned by AI.
Keep reading
Read next
How to defend
A sourced plan for UK organisations to defend against AI-assisted attacks: exposure, patching, payment checks, AI tool limits and breach readiness.
Read the guide
Timeline
A dated, sourced timeline of AI-related cyber attack cases and UK and Five Eyes guidance, from the Arup deepfake fraud to the 2026 call to action.
Read the guide
Questions
Sourced answers to common questions on AI cyber attacks: deepfakes, phishing, faster exploitation, prompt injection and what UK organisations should do.
Read the guideHow CyPro can help
Need a hand with any of this?
Patching is one job. Keeping a vulnerable system safe until it is patched, and proving it, is several more. CyPro's consultants, testers and SOC analysts can take any of them off your plate.
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
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
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
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
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
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
Know your exposure
Not sure you are protected?
AI makes it quicker to find and use weak points, so it helps to find them first. A penetration test shows which of your internet-facing systems, staff processes and AI tools an attacker could actually use.