Imagine this: you’re sitting at home, tired after a long day, and your phone buzzes with a voice note from a local candidate you have actually met at a community meeting. The voice sounds exactly like theirs — the same tone, the same rhythm, the same little pauses and expressions. But the message is strange and urgent. It says that the voting station in your area has been moved at the last minute, or that a respected community leader has secretly switched sides, or that something has happened that could affect the election. The message asks you to spread the word quickly, because people need to know. You feel a knot in your stomach. You don’t want to stay silent if something important is happening. So you forward it to your family group. They forward it to their friends. Within a few hours, thousands of people have heard a lie that sounded exactly like the truth. This is not a distant science fiction scenario. It is a real possibility in South Africa’s local government elections, where artificial intelligence voice cloning, deepfakes and digital impersonation could make election misinformation much harder to detect. What used to require expensive video equipment and technical experts can now be done with freely available tools and a small amount of audio. And the people most at risk are not just politicians or public figures. They are ordinary voters, exactly like you, who just want to make informed decisions.
The Communications Risk Information Centre, known as COMRiC, has issued a warning that should make every voter sit up and listen. The same technologies already being exploited by sophisticated telecommunications fraudsters to manipulate consumers are now powerful enough to be turned against voters. AI-generated voice cloning can reproduce a person’s voice from just a few seconds of audio. Deepfakes can create realistic video footage of people saying things they never said. Spoofed messages can make a harmful text appear to come from a trusted number. Compromised accounts can be used to spread false information from real profiles. Social engineering has become so advanced that criminals and political manipulators can mimic the style, tone, and emotional pressure of a trusted individual in ways that are almost impossible to distinguish from reality. This matters especially at local government level, where information travels quickly through community WhatsApp groups, social media networks and direct messaging. A convincing voice note apparently from a ward candidate, a fraudulent message claiming that a voting station has moved, or an AI-generated recording designed to provoke anger or fear in a particular community could spread within minutes. The danger is not simply that false information can travel quickly. The danger is that technology is making that false information increasingly convincing, and every new tool makes it harder for ordinary people to know what to believe.
We have seen this before in the world of fraud. Criminals have become increasingly adept at impersonating trusted individuals and organisations, exploiting urgency, fear and emotion to persuade people to act before they have had the opportunity to verify what they have been told. In the past, recognising someone’s voice or seeing a familiar name or number on your phone gave you a degree of confidence about who you were dealing with. That assumption is becoming increasingly dangerous. AI can make an impersonation extraordinarily convincing. It can imitate the warmth of a friend, the authority of an official, or the urgency of a community leader. When this technology is used in an election environment, the potential consequences go beyond financial loss. False information can influence behaviour, damage reputations, inflame community tensions and undermine confidence in the electoral process itself. This is not a prediction that such interference will necessarily occur during these elections. It is a warning that the capabilities already exist, and that the techniques used in telecommunications fraud demonstrate how vulnerable we all are to being deceived when someone speaks with a familiar voice or writes from a familiar account. The threat is deeply human, because it is built on trust. And once trust is broken, the entire democratic process becomes weaker.
So what should an ordinary voter do? COMRiC suggests a simple rule whenever you come across unexpected or provocative election-related information: stop, verify and don’t amplify. The natural response to a shocking message is to act quickly, but that is exactly what a manipulator wants. If a message shocks you, frightens you, makes you angry or demands immediate action, that should be the moment you slow down. Do not assume that a familiar number, picture, account or even a familiar voice proves who is behind the communication. None of those things are reliable anymore. Instead, verify the information independently before acting on it or forwarding it. An unexpected or inflammatory claim should be checked against a second, independent source. Information concerning voting stations, registration, voting dates or electoral procedures should always be verified with the Electoral Commission of South Africa. This may feel like a small step, but it is powerful. Every time suspicious content is forwarded, its potential audience grows. You can inadvertently become part of the distribution network for the very disinformation you are trying to question. So pause. Ask yourself: do I know this is true? Could this be a fake? What would happen if I am wrong? A few moments of doubt can prevent a great deal of harm, and a single person choosing to verify instead of share can break the chain.
If a communication appears fraudulent or deliberately misleading, preserve the evidence rather than delete it. Screenshots, telephone numbers, messages and voice notes can assist authorities, telecommunications operators and other relevant organisations in identifying patterns of suspicious activity. Don’t just throw up your hands and hope someone else will deal with it. Report it. Suspected election disinformation should be reported through the appropriate electoral channels, while suspected telecommunications fraud should also be reported to the relevant mobile network operator. Reporting is not only about protecting yourself. It is about protecting your community and helping those who are trying to stop the abuse. The broader challenge will require cooperation between telecommunications operators, financial institutions, technology platforms, electoral authorities, law enforcement agencies and consumers. No single organisation can deal with this problem in isolation. Malicious actors exploit the gaps between systems, institutions and industries. That makes information sharing and collaboration increasingly important. It also means we all have a role to play. Local government elections depend on community trust. When that trust is attacked, the effects can be felt long after the votes are counted. The people who spread disinformation want us to feel isolated and confused, but if we work together — reporting suspicious content, verifying facts, sharing reliable information — we make it much harder for them to succeed.
Ultimately, voters remain an important line of defence. Technology may have changed the sophistication of the threat, but one of our strongest defences remains remarkably simple. Before you believe it, before you act on it, and especially before you share it, verify it. This is not about being suspicious of everyone. It is about being careful with something valuable: our democratic process. It is about remembering that behind every message, voice note or video is a human being trying to make sense of a complicated world. South Africans have already shown great resilience in defending their democracy. This election season, that resilience needs to be extended to the information we choose to trust. We cannot stop every fake from being created, but we can stop it from spreading. We can refuse to be the person who forwards a lie. We can be the person who asks the difficult questions. We can be the person who reminds our friends and neighbours that if something seems designed to make us angry or afraid, it might have been designed precisely for that reason. The technology is new, but the wisdom is old: look before you leap, think before you speak, and verify before you share. If enough of us do that, the manipulators lose their power. And by protecting ourselves from being deceived, we do something even bigger. We protect the trust that makes free and fair elections possible, and we ensure that the voices South Africans hear are the ones they can truly rely on.