Detection
Can deepfakes be detected reliably?
Not reliably enough to be your only control. Detection tools are being tested, and the government expects its results to set industry standards, but the work is not finished.
Last reviewed against Building, Government Office for Science (UK Government) and The Register. Sources
The testing
What the government is doing
The Home Office is working with the Accelerated Capability Environment, DSIT, DCMS, HMRC and the Alan Turing Institute on what it calls a world-first deepfake detection evaluation framework [2]. It tests how well tools identify harmful deepfakes, including those used for impersonation and fraud, and intends the results to set new standards for industry [2]. A separate challenge hosted by Microsoft involves over 350 experts [2].
The practical view
What to do now
Gartner's survey reported that the share of deepfake audio calls causing harm fell from 6% to 2% where an audio screening service was used [7]. That suggests tools can help, though the article gives no sample size [7].
A sensible stance is to treat detection as an extra layer:
- Keep the call-back rule and second approver in place whatever a tool says.
- Ask any supplier what it was tested against and how it handles live calls.
- Record what the tool flags, and review misses.
Related: how to spot a deepfake call and payment approval checks.
The control that does not depend on detection
A rule to ring back on details you already have still holds when a fake goes unnoticed by staff and by software alike [9].
Keep reading
Read next
Questions
Sourced answers on deepfake fraud: how it works, the Arup case, how common it is, how to verify requests and what to do if money has been sent.
Read the guide
Am I exposed?
A practical checklist for UK organisations to see which payments, resets and calls could be hijacked by a cloned voice or fake video call.
Read the guide
How to stop it
An ordered plan to stop deepfake fraud: second-channel checks, call-backs, blame-free reporting, rehearsals and a ready response if money has moved.
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?
Deepfake fraud is stopped by decisions about who may release money or access and how a request is confirmed. A virtual CISO helps leaders set those rules and have them tested, without a full-time hire.