This week, we're turning our Cyber Security Awareness Month series toward one of the fastest-growing threats facing schools and trusts: AI-generated deepfakes, starting with audio and voice cloning scams.

It's not just phishing emails anymore

For years, "spot the scam" meant checking for a dodgy email address, a spelling mistake, or a link that didn't quite match the sender. Cyber criminals have moved on. With freely available AI tools, a scammer now needs only a few seconds of someone's voice, lifted from a school video, a parents' evening recording, a voicemail greeting, or a video on social media, to generate a convincing clone of it.

What is a voice cloning scam?

A voice clone is an AI-generated copy of a real person's voice, built from a small sample of genuine audio. Once trained, it can be used to generate new sentences, in that person's voice, saying things they never said.

In a school setting, this typically turns up as:

• A phone call that sounds exactly like the headteacher or IT support, urgently instructing finance staff to make a payment or share bank details, asking you for your password or PIN.
• A voicemail that sounds like a parent, requesting a change to pick-up arrangements or safeguarding information.
• A voice note circulated on social media or messaging apps, falsely attributed to a member of staff or a pupil

Because the voice sounds right, it bypasses the instinct most of us rely on; "I'd know if it wasn't really them." That instinct is no longer a reliable safeguard.

Why schools are a target

Schools and trusts hold a combination that's attractive to fraudsters: processes for urgent payments (trips, suppliers, emergency repairs), publicly available audio and video of staff (assemblies, promotional videos, parents' evenings, open days, much of it on the school's own website or YouTube channel), and a culture that quite rightly prioritises responsiveness to parents and safeguarding concerns over suspicion. On top of that, staff who are already busy and stressed, who might bypass additional checks due to circumstances. Criminals exploit exactly that responsiveness.

Why the risk is growing, and growing fast

Three things are converging to make this a genuine, rising risk for the education sector rather than a theoretical one:

• The barrier to entry has collapsed. Voice cloning tools that once required technical skill and expensive software are now free, fast, and require only a few seconds of audio to produce a convincing result.
• Schools generate a lot of exploitable audio. Assemblies, promotional videos, open days, parents' evenings, podcasts, and staff appearing in local news or on social media all give an attacker raw material, often without anyone realising it's there.
• The financial and safeguarding stakes are high. Schools and trusts regularly handle urgent payment requests (trip deposits, supplier invoices, emergency repairs) and urgent safeguarding communications,  exactly the categories of request an attacker wants to impersonate, and exactly the categories where staff are trained to respond quickly rather than question first.

Put together, this means the risk isn't just "could this happen to us", it's "when it happens, will anyone have been told what to look and listen for." For most schools right now, the honest answer is no. Voice cloning doesn't yet feature in most cyber awareness training, which means staff are being asked to defend against a threat they've never been briefed on.

This also overlaps with a risk we've flagged before: Shadow AI:  the unauthorised, unmonitored use of AI tools by staff or others,  which can make it harder to spot when AI has been used against the organisation, not just within it.

Building this into your training, not just your awareness

Spotting a voice clone isn't really a technical skill, it's a habit, and habits need to be trained in, not just mentioned once in an assembly or an email. A few practical additions worth making to your existing cyber security training and induction:

• Name the threat explicitly. Don't fold voice cloning into generic "be alert to phishing" messaging, most staff have never heard that this is possible, so it needs its own line in your training, not a footnote.
• Rehearse the response, not just the awareness. Staff should practise what they'd actually do if a call like this came in: who they'd contact, on what number, and what they would (and wouldn't) action in the meantime, in the same way fire drills rehearse a response rather than just explain a risk.
• Target the right roles first. Finance, senior leadership's PAs and office staff, and anyone who can authorise a payment or release personal data over the phone are the highest-value targets for this specific scam, and should be the first group trained. Although be aware, everyone should be vigilant about receiving calls from anyone claiming to be IT support; ensure there is a protocol in place.
• Refresh it regularly. The tools behind this threat are improving quickly, so this can't be a one-off training slide, it needs revisiting at least annually, alongside your other cyber security refreshers.
• Make reporting easy and blame-free. Staff who almost acted on a cloned voice, or who did act and realised afterwards, need a quick, non-punitive way to flag it ; the faster it's reported, the faster a payment can potentially be recalled or an account locked down.

💡Your Daily Cyber Tip: verify using a separate, trusted method

If you receive an urgent request by phone or voice note, especially one involving money, personal data, or a change to safeguarding arrangements, stop and verify it through a separate, trusted channel before acting. Call the person back on a known number (not one given to you in the suspicious message), or confirm face-to-face. A genuine, urgent request will always survive a two-minute call back.

Where this sits in the DfE Digital & Technology Standards

Deepfake and voice cloning threats fall under the Cyber Security standard, specifically cyber awareness and user vigilance, and connect directly to the Digital Leadership & Governance standard, which expects schools to have clear escalation routes when something doesn't look (or sound) right. Building AI literacy into your cyber awareness training, as set out in the DfE's standards, is now as important as spotting a phishing email.

Further reading from our AI Best Practice Library

If you haven't already, it's worth reviewing DPE's dedicated AI resources, which sit alongside this series in our Knowledge Bank:

• AI Best Practice Area: our central hub for AI guidance, checklists and policy templates
• AI Policy Template: recently updated for 2026, covering accountability, transparency and data minimisation
• Is your organisation at risk from Shadow AI?: unauthorised AI use and why it matters
• AI in Schools: Six Pillars for Getting it Right: our Serious Privacy feature with Ralph T O'Brien
• Can you use AI safely in schools? — on keeping AI tool use compliant with data protection law