Scams and social engineering

AI phishing and impersonation: why the old tells fail

Phishing was already the biggest everyday threat to UK organisations. AI makes the messages cheaper to produce and harder to judge by eye, so the defence has to be process rather than instinct.

Last reviewed against NCSC (UK), Anthropic and UK Government (DSIT and Home Office). Sources

Brick firewall with a flame in front of it

The baseline

Phishing before AI

The 2025/2026 government survey ranks phishing first among attack types, at 38% of businesses and 25% of charities [6]. It was also named the most disruptive by 69% of those that had a breach or attack [6]. The Register's report on the survey said about 85% of businesses that reported a breach or attack said it involved phishing [7]. The two figures measure different groups, so quote each one with its source and do not set them side by side as if they were the same measure.

What AI changes

Three effects to plan for

Factory site linked by wireless to a phone with a fingerprint and shield

Polish at scale

The NCSC expects AI to keep making parts of an intrusion quicker and more effective [1]. For inboxes, the practical result is that spelling and tone stop being useful signals.

Tall blue server rack with glowing status lights

Wider access

The NCSC expects AI-enabled cyber tools to put intrusion within reach of more people [1].

Wi-Fi signal above a hexagonal padlock

Voice and video

Synthetic voices and images were used in the Arup fraud [8]. See deepfake video call fraud for how that worked.

Controls

What holds up

The survey does not test which controls stop phishing that uses these tools, so these are process measures rather than measured results. Choose the ones that fit your organisation:

  • A second-channel check for any request involving money, credentials or data
  • One obvious button or address to report a suspicious message
  • A clear statement from leaders that pausing a request is welcome
  • Limits on what a single approver can release alone

Next: the defence plan.

Why process beats instinct

At Arup, the faces and voices on the call were what overcame the employee's early doubts [8]. A check that does not depend on spotting a fake still works when the fake is good.

How 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.

Scanner checking internet-facing systems for weaknesses

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
Cyber security consultants planning hardening work

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
Security analysts monitoring systems around the clock

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 tester probing an application for weaknesses

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 responders containing an attack

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
Security lead advising a board

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
Rocket launching from a cloud

Know your exposure

Not sure you are protected?

Many of the risks here sit in how an organisation uses AI: what tools staff run, what they can reach and who owns the risk. AI governance work sets that out so controls match the tools in use.