The THURSDAYBRIEFING: When the Algorithm Decides What We See: Who Governs Africa's Digital Democracy?

 By Dr Blessing Ivan Vava

On Monday, I joined a discussion on ZiFM Stereo, a leading privately owned radio station in Zimbabwe, about Zambia’s general election and what it revealed about the changing nature of democracy in our region. We spoke about the smartphone as a new political arena, the growing threat of AI-generated misinformation and, ultimately, the deeper question of digital sovereignty.

One argument kept returning throughout that conversation: information no longer simply surrounds an election. It has become part of the infrastructure through which elections are experienced, contested and understood.

A few days ago I came across a story from Australia that at first sounded like a technical policy debate. The Australian government is proposing to give social-media users something deceptively simple: choice over how their feeds are generated. Under the proposed rules, part of the government's “My Feed, My Way” initiative, users would be able to opt out of personalised algorithmic feeds and instead see content largely from people and accounts they have deliberately chosen to follow, part of a broader “digital duty of care” framework that would make technology companies more responsible for foreseeable harms arising from their platforms and algorithms.

This might look like a distant Australian problem. It is not. For countries such as Zimbabwe, Zambia and many others across Africa, the question behind Australia's proposal is becoming increasingly urgent: who decides what we see when we go online and, perhaps more importantly, who decides what we don't see?

Australia is not acting alone. The European Union (EU) has already moved further down this road through its Digital Services Act, which requires the largest platforms to assess and mitigate systemic risks to electoral processes and civic discourse; the EU's election guidelines specifically address recommender systems, generative AI, political advertising, data access for researchers and cooperation with civil society. The significance is not that Europe has found a perfect model; rather, it is that algorithmic governance is increasingly being treated as a democratic-governance issue rather than simply a technology issue.

The Phone in Your Hand Is No Longer Just a Phone

Consider how many of us now experience politics. You wake up and reach for your phone. Before breakfast, you may have received several WhatsApp messages about a politician, opened Facebook to a video about the economy, scrolled past a political clip on TikTok, received a screenshot claiming something dramatic has happened, and found yet another version of the story on X. By lunchtime, you may have encountered dozens of competing versions of reality, but you did not necessarily choose all of this information. The platform chose much of it for you. The smartphone has democratised political communication, but, just as powerfully, it has democratised political manipulation.

An algorithm decided that a particular video might keep you watching. Another post was pushed because it generated strong reactions. Something controversial appeared repeatedly because people kept commenting on it, while a quieter, more nuanced explanation disappeared beneath the noise. The algorithm does not know whether you are looking for truth, it only knows that you are looking. That distinction matters enormously for democracy.

We Saw This in Zambia

My recent work around Zambia's 2026 elections reinforced something I had already been arguing in my writing on digital power and democracy: the smartphone has become part of the political infrastructure of African elections. The campaign now takes place beyond just rallies, television, and newspapers. It happens in WhatsApp groups, on TikTok, through Facebook videos, X posts, and involves influencers, commentators, anonymous accounts, and increasingly advanced digital strategies.

During the Zambian election period, I saw first-hand how quickly political narratives could travel from one platform to another: a rumour became a WhatsApp voice note, the voice note became a Facebook post, the post became a TikTok video, and the video was then presented as “evidence” somewhere else. By the time anyone tried to establish what was actually true, the original claim had already acquired a life of its own. This is where the algorithm becomes politically important, it does not need to create the lie, it only needs to help the lie travel faster than the truth. That is the problem we need to start confronting.

What made this dangerous was not necessarily any single falsehood, but their cumulative effect. Fake GEN-20 result forms circulated as though they were official; a fabricated communication claimed State House had ordered an internet shutdown; and another false claim suggested that UN observers had congratulated a particular candidate. None needed to convince everyone. Their power lay in collectively weakening the shared informational foundation on which a credible election depends. When citizens can no longer agree on what is authentic, who said what, or even which documents are real, misinformation becomes more than a problem of false content; rather, it becomes a problem of democratic trust.

Zambia also offered a glimpse of a harder challenge. A disputed video involving Tonse Alliance presidential affairs chairman Binwell Mpundu circulated online before being deleted, accompanied by competing claims about whether it had been hacked, manipulated, taken out of context or generated using artificial intelligence. Whatever the truth of that particular case, it illustrated what researchers call the “liar’s dividend”: once convincing synthetic media exists, fake material can be presented as real while genuine material can be dismissed simply by saying, “That’s AI.” The democratic danger therefore runs in both directions. AI can make falsehood easier to manufacture, but it can also make truth harder to prove.

Zambia is not unique. Taiwan's 2024 presidential election offered an earlier glimpse of the same problem: false claims of vote-counting fraud circulated through YouTube and TikTok, were amplified by influencers, and were subsequently forwarded across LINE and Facebook to millions of viewers. The lesson is uncomfortable but important, because once an unverified claim acquires visual form and digital momentum, correction becomes much harder than circulation.

Africa's Digital Challenge Is Changing

For years, our conversation about technology in Africa centred on the digital divide, mainly on who has internet access, a smartphone, affordable data, and broadband. Those questions remain important, but I increasingly believe they are no longer enough. A person can have a smartphone, affordable data and access to the world's largest digital platforms and still have very little control over the information environment in which they live. This is why I have argued that Africa's digital challenge is no longer only the digital divide; it is the governance divide. The question is no longer simply whether Africans are connected, it is who governs the systems to which we are connected.

Zimbabwe Should Pay Particular Attention

Zimbabwe has a complicated history with digital governance. The country has experienced internet disruptions, restrictions on digital platforms and growing debates around cyber regulation, and digital-rights organisations have repeatedly warned that laws ostensibly designed to address cybercrime, online harms or national security can also chill legitimate expression. This makes the Australian debate particularly relevant but also particularly delicate. Zimbabwe should learn from Australia's recognition that technology companies cannot simply claim to provide the platform while bearing no responsibility for what happens on it, but it should not respond by giving the state unlimited authority to decide what citizens are allowed to see, share or discuss. The answer has to be democratic digital governance.

What Zimbabwe Could Learn From Australia

First, citizens need meaningful choice. The Australian proposal treats users as citizens, not just consumers: if you don't want an engagement-driven feed built to keep you scrolling, you should be able to opt out. Zimbabwe and other African countries should ask whether their own citizens deserve that same right to understand, and to control, how platforms curate what they see.

Choice alone is not enough. Platforms must answer for the systems they build. We regulate banks because their failures can wreck economies, pharmaceutical companies because unsafe products can kill, broadcasters because they shape public opinion at scale. Platforms that now shape political communication on a similar scale cannot claim exemption. This is not about policing what is offensive or controversial; it is about forcing transparency: how these systems work, what they amplify, and what risks they create.

Election observation should go digital. While observers still monitor voter registration, polling stations, ballot counting, and results transmission, these activities are no longer enough since a large part of campaigning now happens online. Brazil shows what this can look like: ahead of its 2026 elections, its electoral authorities moved beyond traditional campaign rules to confront AI-generated political content head-on. African observers should be doing the same, tracking political advertising and targeting, algorithmic amplification, misinformation networks, AI-generated content, uneven enforcement of platform rules, and how political content in African languages is treated. The digital campaign is now part of the election. Observation has to follow it there.

Citizens need algorithmic literacy, not just internet access. While we have invested years in teaching people how to get online, it's now essential to educate them about how the internet influences their lives. A video that keeps appearing on a young person's TikTok feed might not be the most accurate one. It could simply be the video that the system predicts will keep their attention the longest.

Regulation needs to be balanced. A framework designed to control Big Tech can also be used to abuse power by Big Government. Zimbabwe cannot swap one problem for another. A proper African digital governance framework should prioritise freedom of expression, privacy, transparency, due process, and independent oversight. No regulator should be allowed to dictate what is true.

This Is Ultimately About Power

My research on China's increasing technological presence in Africa inspired me to consider technology through a lens I call geosociotechnopolitics, which explores the connections between geography, society, technology, and political power. The algorithm is now part of that equation. We often think about technological sovereignty in terms of fibre-optic cables, data centres, cloud infrastructure and who owns the technology. But there is another sovereignty we have paid far less attention to: the sovereignty of attention. Who decides what Africans see? Who determines what is amplified or buried? Who profits from our attention, and who bears the democratic consequences? These questions matter because attention is becoming one of the most valuable resources in contemporary politics.

Africa Must Not Simply Import Other People's Digital Futures

Australia, the EU, Brazil and Taiwan offer different lessons about user choice, platform accountability, AI regulation and misinformation. Zambia offers an African illustration of how these same pressures are already reshaping electoral politics. Africa's goal should not be to imitate others. Instead, it must define what democratic digital governance looks like based on our unique political, social, and institutional contexts, well before the next election.

Our media systems, political institutions, levels of connectivity and histories of state power are all different from those of Australia, the EU, Brazil or Taiwan. The danger is that we import technology from one part of the world, regulation from another and political assumptions from somewhere else, and end up with a digital governance model that belongs to nobody except the companies and governments powerful enough to shape it.

Africa needs to become a rule-maker, not merely a rule-taker, in the digital age. Regional institutions such as SADC and the African Union should discuss common principles for algorithmic transparency, political advertising, AI-generated election content, platform accountability, and digital rights. Electoral commissions should develop permanent digital-election monitoring capacities rather than waiting until election week; civil society should demand access to platform data and greater transparency; universities should build African expertise in algorithmic governance; and governments should resist the temptation to use “misinformation” as a convenient justification for silencing political opponents.

Governance is necessary, but it is not enough. Africa cannot simply regulate algorithms built elsewhere; it must also develop the capacity to build its own. AI systems rooted in African languages, data and political and cultural realities are not a luxury for the future. They are part of what genuine digital sovereignty should mean. A continent that only consumes and regulates technologies developed elsewhere will continue to negotiate from a position of weakness. Building African technological and algorithmic capacity through our universities, private sector and public institutions must therefore sit alongside the debate about digital rights, democracy and sovereignty. Ultimately, these are not separate conversations. They are part of the same struggle over who gets to shape Africa’s digital future.

The Real Democratic Question

The Australian debate has forced an uncomfortable question into the open. Perhaps the biggest threat to democracy in the digital age is not simply that someone can post something false. It is that a small number of private companies have acquired extraordinary power to determine what millions of people see, how often they see it, and how long it remains visible. That is a political, economic, and technological question, but above all, it is a democratic question that Zimbabwe and Africa should be asking now.

The next election will not simply be fought between political parties. It will also be fought inside digital systems that most voters, and many politicians, barely understand. The ballot paper may still be printed in a government office, and the polling station may still be a school or community hall, but the political reality surrounding that voter may have been shaped for months by software sitting on a server thousands of kilometres away.

That is why Australia's proposed reforms matter to us, not because Australia has all the answers, but because it has recognised something Africa can no longer afford to ignore: the algorithm has become part of the public sphere. And once algorithms become part of democracy, democratic governance must also confront the power of algorithms.

The question for Africa is therefore no longer whether technology will shape democracy. It already does. The question is whether our democratic institutions, citizens and societies will have any meaningful power to shape the technology in return. Africa needs neither Big Tech democracy nor Big Government democracy. It needs citizen-centred digital democracy. Ends

About the Author

Dr Blessing Ivan Vava is a Zimbabwean communication scholar, researcher and governance practitioner whose work focuses on the intersection of technology, power, democracy and society in Africa. He holds a doctorate in Communication Studies from the University of Johannesburg, where his research examined Chinese information and communication technologies in Zimbabwe and developed geosociotechnopolitics as a framework for understanding the relationship between technology, society and political power.  He writes The Thursday Briefing, a commentary series exploring digital governance, democracy, artificial intelligence, geopolitics and Africa’s place in the changing global technological order.

He can be contacted at: blessingvava@gmail.com