Last month, news broke of a Canadian fraud ring that used AI to clone the voices of elderly Americans’ grandchildren. Between 2021 and 2024, scammers called seniors across 46 states, pleading for urgent bail money or medical fees, each ruse delivered in a perfect imitation of a cherished grandchild.
By the time police shut down the operation, victims had wired more than US$21 million to the fraudsters. If this chilling scam could thrive in a high tech country, what stands between Zimbabwean families and the next wave of AI voice cloning?
Here in Harare, Bulawayo, and even small villages along the Gweru road, mobile phones are the lifeline of kinship. We call our children upcountry, check on relatives abroad and stay connected to the diaspora in South Africa, the UK or Australia. When a familiar voice crackles over the line, our hearts soften. We trust voices in a way we rarely trust texts or emails. But AI voice cloning has turned our most sacred trust into a weapon.
Imagine a pensioner in Mbare receiving a frantic call at 10pm. A weeping voice claims to be his grandson, arrested on trumped up charges, begging for cash to bail him out. The grandson’s tone is panicked, precise and unmistakable. The grandfather, single and lonely, rushes to send US$100 via EcoCash or One Money, hoping to save his only grandchild. Only later does he learn the voice was a digital fabrication. By then, the money is gone, along with any hope of tracing it.
This is not mere financial fraud. It is the emotional hijacking of human relationships through synthetic media. AI voice cloning tools have become startlingly accessible: a few minutes of audio, an open source script and a modest laptop suffice to produce near perfect impersonations. Scammers press a few buttons, generate a targeted message and dial thousands of numbers in automated bursts. The victims, often isolated, trusting and unaware, answer out of love, not caution.
Zimbabwe has long grappled with mobile money and lottery scams, but AI voice cloning raises the stakes exponentially. In traditional vishing (voice phishing), fraudsters rely on clumsy accents or pattern dialling dead air; alert users often smell a rat. With AI, those clues vanish.
A cloned voice can mimic accents, pace and emotion. Even tech-savvy citizens can be fooled when they hear “Melody” begging for funds. A single deepfake call is all it takes to break down defences built over the years.
What if we could stop these scams before they strike? The Responsible AI Institute proposes “RAISE pathways”, a set of safeguards designed to protect the public from AI-driven impostors. Zimbabwe must adopt similar measures, tailored to our unique telecom landscape and social fabric. At the heart of RAISE lie three pillars: real time verification, secure identity protocols and voice provenance tracing. Each of these can, and should, be woven into our national security strategy.
- Byo armed robber in court
- Woman jailed 12 years for indecent assault
- Magistrate jailed 3 years for abuse of office
- Zim’s poor batting hands India series
Keep Reading
First, let’s look at real time verification tools. In simple terms, every call must carry a digital stamp of authenticity. Just as banknotes bear watermarks to prove their legitimacy, voice calls can be “watermarked” by telecom networks. When a subscriber calls, the network adds an encrypted marker that devices or caller-ID services can validate. If the marker is missing or fails verification, the recipient sees a clear warning: “Caller identity unverified.” This would instantly alert users that something is amiss.
Second, secure identity protocols. Phone calls alone should never be the sole proof of identity. We need “code-word” systems and call back procedures. For example, every family could agree on a unique passphrase, known only to close relatives, which must be spoken at the start of any money request call.
Alternatively, calls could default to a two-step verification: the caller dials, states their request, and the recipient’s phone automatically hangs up and calls back on a different channel. If the scammer cannot respond on that authenticated line, the ruse fails.
Third - voice provenance tracing. Advances in audio hardening allow service providers to embed inaudible markers in legitimate voice streams. These markers carry metadata, date, time, originating cell tower and network operator, which can be extracted and cross checked. When law enforcement seizes recordings, analysts can trace the call’s true origin rather than following money through anonymous accounts. Over time, repeated breaches from the same spoofed line raise alarms and trigger automatic network blacklisting.
Implementing these safeguards requires coordinated action. Zimbabwe’s Postal and Telecommunications Regulatory Authority (Potraz) must lead the charge, issuing guidelines that compel network operators and handset manufacturers to adopt authentication standards. Potraz can mandate “secure calling” protocols to safeguard citizens against AI fraud.
Network operators, Econet, NetOne and Telecel, should retrofit their switching centres with AI-aware signalling systems. These systems would inspect call-setup messages in real time, inject digital authenticity markers and flag anomalies
Handset makers, in turn, must ensure that default dialler apps can read and display verification statuses prominently on every incoming call. Phone users should see green icons for verified calls and red alerts for unverified ones.
Beyond technology, legal frameworks must keep pace.
Zimbabwe’s Financial Intelligence Unit and police cybercrime units should classify AI voice cloning as an aggravated fraud offence, subject to harsher penalties than traditional scams. The courts can send a strong deterrent message by imposing jail sentences and hefty fines on convicted fraudsters. At the same time, law enforcement agencies need specialised training in digital forensics to extract provenance data, analyse cloned audio samples and track down the criminals who hide behind bytes.
Public education is equally vital. Many Zimbabweans remain unaware that a familiar voice over the phone is no longer proof of identity. Community radio stations, church gatherings and village meetings should run programmes explaining how AI voice scams work and teaching preventive habits: never transfer money on first demand, always demand an agreed code word and verify urgent requests through alternative channels—such as visiting a relative in person or video calling over WhatsApp.
Urban and rural alike can benefit from simple checklists. A poster in clinics and district councils might read: “Scam Alert: If a relative calls asking for cash, pause and verify. Ask for your family’s secret word.
If in doubt, hang up and call back on a known number.” By repeating these messages in Shona, Ndebele and English, we reinforce them across linguistic divides. Local champions, school teachers, agricultural extension officers, and village heads can amplify the warnings.
We must also equip vulnerable demographics with direct support lines. Senior citizen helpdesks, staffed by volunteers and connected to telecom customer care teams, can offer immediate advice to distressed callers.
If an elderly person suspects a scam, a single call to the helpdesk could trigger a real time investigation: the helpdesk operator pings the network to confirm whether the incoming call bore a valid authenticity marker. Within seconds, the operator can advise whether to proceed or to report the incident.
Private sector partners have a role too. Banks and mobile money operators can integrate scam detection alerts into their apps. If a user attempts to send an unusually large amount to a new beneficiary, the app could prompt: “Is this a payment to a verified family contact? If you received a voice call requesting this transfer, press ‘No’ to pause the transaction and call your helpdesk.” Such friction, far from being an annoyance, can save livelihoods.
Critics may argue that these measures are too costly or technically complex. But Zimbabwe has shown its ability to leapfrog technologies before: our mobile money revolution, for example, far outstripped traditional banking in accessibility. The same spirit of innovation can drive secure calling adoption.
Sagomba is a chartered marketer, policy researcher, AI governance and policy consultant, ethics of war and peace research consultant. — Email: esagomba@gmail.com, LinkedIn: @Dr. Evans Sagomba, X: @esagomba.
