Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Voice cloning fraud uses a synthetic copy of someone’s voice, built from seconds of audio, to authorise payments, reset credentials, or pressure a relative into sending money.

What Is Voice Cloning Fraud?
Voice cloning fraud is impersonation using synthetic speech. A few seconds of recorded audio—from a conference talk, a podcast, a voicemail greeting, a social media clip—is enough for current tools to produce convincing speech in that person’s voice, in real time, with the right accent and cadence. The fraud that follows is familiar: a payment that must be made today, a credential that must be reset urgently, a family member in trouble. What has changed is that the voice on the line no longer counts as evidence of who is speaking.
The first widely reported incident came in 2019, when the chief executive of a UK energy subsidiary transferred around €220,000 after a call from what he believed was his German parent company’s chief executive, complete with the expected accent. The scale rose sharply afterwards. In 2024, the engineering firm Arup confirmed that an employee in its Hong Kong office had transferred approximately HK$200 million—roughly £20 million—after a video conference in which every other participant was a synthetic recreation of a colleague. Consumer-facing variants are now routine: UK banks and Action Fraud have both warned about calls in which a cloned voice of a family member asks for emergency funds, and the technique has appeared in impersonation of bank fraud teams.
Three processes carry most of the risk. Payment authorisation, where a call confirms an instruction received by email, is the classic business email compromise amplifier. The IT service desk is the second: a caller who sounds like a known employee requesting an MFA reset or a password change defeats a process built on recognising a colleague, and this route has featured in several high-profile intrusions. Third, voice biometrics used for customer authentication in banking and telecoms has been demonstrated as defeatable by synthetic speech, which is why providers have moved towards liveness detection and treat voice as one factor rather than as identity.
Remove voice from the decision. For payments and bank detail changes, verify through a separate channel using contact details held in your own records, never a number supplied in the request, and require dual approval above a threshold that seniority cannot waive. For the service desk, replace voice recognition with something verifiable: a push approval to an enrolled device, a manager confirmation through a tracked ticket, or in-person verification for high-impact resets. Some organisations agree a code phrase for executive requests, which is cheap and surprisingly effective. Tell staff explicitly that no legitimate request will ever require them to bypass a process because of urgency, and that declining to act on a call is always acceptable. Detection tools for synthetic speech exist but are not dependable enough to be the control, and provenance signalling addresses recorded media rather than live calls.
Report attempted and successful fraud to Action Fraud, and notify your bank immediately—UK reimbursement rules for authorised push payment fraud make prompt reporting and a documented verification process commercially significant. Where personal data was disclosed during the call, assess it as a breach under UK GDPR. There is no specific UK offence of voice cloning, but using a clone to obtain money engages the Fraud Act 2006, and the Online Safety Act 2023 has extended protections around certain synthetic content. At Privia Security, we run authorised vishing exercises against agreed scenarios, including service desk and finance workflows, as part of our Social Engineering Services.
Modern tools produce a usable clone from a few seconds of clear speech, and better results from a minute or two. A conference recording, a podcast appearance, a voicemail greeting, or a short video clip is sufficient. Assume that anyone who has spoken publicly, and most people who have left a voicemail, can be cloned convincingly enough for a phone call.
Not reliably by ear, and detection tools are not yet dependable enough to act as a control, particularly over compressed telephone audio. Asking unexpected questions helps less than it used to because real-time systems keep up. The practical answer is procedural: verify through a second channel rather than trying to judge authenticity in the moment.
Treat it as one factor, not as identity. Synthetic speech has been shown to defeat voice matching, which is why providers have added liveness detection and behavioural signals. For anything high-impact, pair it with a device-based factor or a separate verification step, and never let a voice match alone authorise a payment or a credential reset.
Stop treating a recognised voice as identification. Use a push approval to an enrolled device, a confirmation from the requester’s manager through a tracked ticket, or in-person verification for high-impact actions such as MFA resets and password changes. Record the verification method used on every such ticket, so both compliance and investigation have something to work from.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.