How Zambia’s 2026 election showed what happens to an information landscape when synthetic content becomes ordinary
Guest Op-ed : Luckson Mulako Sikananu – Learning Experience Manager at RNW Media/RNTC
Elections concentrate everything. Power, fear, identity, the future; all compressed into a narrow window where the outcome feels existential, and the rules feel temporary. That compression creates an enormous incentive to deceive. Not because people are inherently dishonest, but because the stakes reframe deception as justified. A fabricated claim that shifts a few thousand votes is not experienced by its creator as a lie. It is experienced as a necessary intervention.
This is not new. What is new is that the tools for acting on that incentive have become effectively free. Generative AI has collapsed the cost of producing convincing synthetic content; fake videos, fabricated official graphics, forged institutional statements, to near zero, precisely at the moment when the information environments in which they circulate are least equipped to absorb them. Research from the Africa Center for Strategic Studies has warned consistently that African elections are increasingly targeted by digitally manipulated content designed to suppress turnout, delegitimise institutions, and manufacture false narratives of international endorsement or condemnation.
Zambia’s August 2026 general election offered a concentrated example.
During the election period, I saw it firsthand, on a friend’s WhatsApp story. An image that looked like a front page of the Zimbabwe Herald, the right masthead, the right typographic weight, a headline announcing that Zimbabwe’s president, Emmerson Mnangagwe, had denounced Zambia’s elections. And a Facebook page styling itself as a news outlet shared it. It was not a Herald front page. No such headline was published; no such statement was made. It did not need to go viral. It only needed to be reshared enough: from a page to a timeline, from a timeline to a group, from a group to a WhatsApp story, for the claim to settle into someone’s understanding of events as though it had always been there.
This is the thing worth paying attention to, and it is easy to miss if the conversation stays fixed on deepfakes. The most effective synthetic content in Zambia’s election was not the most technologically impressive. It was the content that borrowed credibility from institutions that still have some.
Zambia’s electoral period produced its share of high-production forgeries. A fabricated video showing the Electoral Commission chairperson under arrest. An AI-generated clip depicting deceased former president, Edgar Lungu, in circumstances designed to provoke grief. On the eve of the vote, the Zambian government flagged a coordinated plan to release synthetic endorsements attributed to figures who could not contradict them. After polls closed, fake social posts carrying forged UN emblems claimed international observers had congratulated a candidate before results existed.
But alongside these sat a quieter, more effective category: fake mastheads, fabricated wire-style graphics, screenshots of statements never made. Cheap to produce, difficult to trace, and calibrated not to persuade through spectacle but to slot seamlessly into a scroll. Their power lies in mimicry of form. They do not ask you to believe in something extraordinary. They ask you to recognise something familiar.
The trust does not disappear. It gets counterfeited.
Zambia’s traditional media has thinned considerably. Commercial pressure, closures, contraction of newsroom capacity, and a broader migration of attention have left legacy outlets holding less of the public’s day-to-day information diet than they did a decade ago. The assumption is often that this authority simply evaporates.
It does not. It gets appropriated.
The visual grammar of a newspaper front page still signals seriousness to a Zambian reader, even to one who has not bought a physical paper in years. The formatting of an official statement still carries institutional weight. What synthetic content does is decouple that signal from the institution that earned it. The masthead becomes a costume.
This is why the standard framing of “verification is getting harder” is only half the picture. Verification is harder, yes. But the deeper damage is that the existence of convincing fakes contaminates authentic material too. Once a fabricated video of the Electoral Commission of Zambia (ECZ) chairperson circulates widely, every genuine communication from that office arrives carrying a shadow. The forgery does not need to be believed to be effective. It only needs to make belief itself feel unsafe.
Distribution is the actual story
None of this would matter at scale if the distribution architecture had not changed underneath it.
The dominant vehicles in Zambia are now individual Facebook pages with followings that rival established outlets, and, more consequentially, WhatsApp channels and groups. This matters for a reason that goes beyond reach.
Content on WhatsApp does not arrive as broadcast. It arrives as personal transmission. It comes from a cousin, a workmate, a church group, a former classmate. The social relationship performs the authentication that institutional verification used to. And it lands in a closed space, where there is no counter-signal, no comment section pushback, no visible correction, and albeit critically, nothing that external fact-checkers can even observe, let alone respond to.
Zambia has real information integrity infrastructure. iVerify Zambia, relaunched ahead of this cycle with UNDP and Panos Institute Southern Africa, actively debunked significant falsehoods, including the fabricated ECZ arrest video. This work matters. But fact-checking is structurally reactive, and it operates almost entirely in the open web while the most consequential circulation happens in encrypted, invisible spaces. By the time a correction is published, the claim has completed its journey through networks the correction cannot enter.
On audiences
Somewhere, there is a lazy conclusion to be made here, which is that people are simply too credulous. This assertion should be resisted.
Susceptibility is not a character flaw. It is a rational response to a degraded environment. When institutional verification is weak, when the visual cues you were taught to trust can be forged for the cost of nothing, when authoritative sources are less present in daily life than they once were, and when information arrives through a relationship rather than a channel, believing what reaches you is not stupidity. It is the only available heuristic.
Media literacy work is essential, but framing it as a fix for individual gullibility misdiagnoses the problem. The task is not to make people more suspicious. Universal suspicion produces disengagement, which serves manipulation just as well as credulity does. The task is to rebuild the availability of trustworthy signals, so that scepticism has somewhere to land.
But even that is not enough on its own. The harder work is understanding narrative infrastructure itself; who is producing this content, what outcome it is designed to serve, and through what channels it reaches the people it is meant to move. In Zambia, some of the synthetic content was clearly partisan, engineered to shift voter behaviour in the final hours of the elections. Some appeared designed not to persuade but to confuse, to flood the space with competing claims until no single account of events could hold. And some of it arrived from outside Zambia entirely, part of a broader pattern in which elections across the Global South become testing grounds for influence techniques that are refined and redeployed elsewhere.
This is not a Zambian problem with Zambian solutions. It is a global production chain with local points of impact. Responding to it means looking upstream; at the platforms that host it, the economic models that incentivise it, and the geopolitical interests that fund it, not only at the audiences who receive it.
What this asks of us
Zambia is not exceptional. It is early. The conditions that made this cycle what it was; thinning legacy media, generative tools at effectively zero cost, encrypted distribution, and a public reasonably unsure what to trust, are present in most contexts across the world.
Information integrity cannot be a matter of catching falsehoods faster. The forgeries will always be cheaper and quicker than the corrections. The work has to move upstream: strengthening the institutions whose credibility is being counterfeited, building newsroom protocols before the crisis rather than during it, and taking seriously that the most important information spaces are now ones we cannot see into.
The counterfeiters understand exactly what they are stealing. The question is whether the institutions being impersonated understand what they still have.
Disclaimer: The views expressed in this guest essay are those of the author. RNW Media is not responsible for the opinions shared.