Manufactured Outrage: How Disinformation Networks Exploit South Africa’s Anti-Foreigner Sentiment
Imagine scrolling through your social media feed and encountering a video that makes your blood boil. Red emojis flash across the screen. The caption warns of a crisis. The comments are angry. Within minutes, hundreds of people have shared it. You watch. You react. You share. But before doing so, there is another question worth asking: Why did this content reach you in the first place? That question is central to new research by Tales of Turning, a South African non-profit organisation that uses social media and online conversations to prevent violence, counter hate speech and help communities heal. Its work is rooted in the understanding that words can incite harm, and that the digital space is now a frontline for conflict prevention. The organisation analysed a randomised sample of publicly available social media posts between 30 May and 30 June 2026, using artificial intelligence alongside human analysis to identify patterns of misinformation, disinformation, hate speech and coordinated amplification. Its findings are sobering. The report found that 27% of the anti-foreign-national narrative analysed was linked to a disinformation network of disruptor and bot accounts. In other words, more than a quarter of the hostile content targeting foreign nationals in the sample was not simply a spontaneous expression of public anger. It was amplified by coordinated actors. The findings do not suggest that all criticism of immigration policy is manufactured. They suggest that a significant portion of the online conversation is being shaped by actors seeking to inflame tensions. The observation raises a difficult question about the online information environment: how much of the outrage people encounter is organic, and how much is being deliberately amplified? The answer matters because online narratives do not stay online. They shape how people view their neighbours, how they vote, how they treat strangers and, in the worst cases, whether they take to the streets. Understanding the machinery behind viral anger is not about dismissing genuine grievances. It is about recognising that some forces benefit from division and have learned to exploit the very platforms designed to connect us. The platforms that allow people to organise for social justice are the same platforms that allow hate to spread. The question is whether society can develop the tools to tell the difference. The video may be old, out of context, or even fabricated. But by the time that is established, the damage has been done. The research is part of a broader effort to understand how social media is used as a tool of social division in South Africa.
Disinformation campaigns of this kind often use divide-and-rule strategies. Their purpose is not to persuade but to polarise. They work by eroding social trust and forcing people into opposing sides. They present one group as a serious threat to another, making it harder for people to find common ground or consider different perspectives. The aim is to turn disagreement into fear and hostility, encouraging people to see those on the other side as enemies rather than as fellow citizens with different views. Tales of Turning’s research identified a recurring narrative in which South Africans are portrayed as citizens, taxpayers and patriots under threat, while foreign nationals are characterised as illegal invaders, criminals or economic threats. That framing is often connected to genuine social problems: unemployment, crime, housing shortages, pressure on public services, failing infrastructure and difficulties faced by small-scale businesses. These problems are real. But they are also complex. Unemployment, for example, is driven by structural economic factors, educational inequality and global shifts, not simply by the presence of foreign nationals. Crime has many causes, including poverty, inequality and weak institutions. Housing shortages reflect decades of planning failures and spatial apartheid. Reducing these complex challenges to a single explanation, that foreign nationals are responsible, can turn legitimate frustration into hostility towards an entire group. It also offers a dangerously simple answer to questions that demand careful, evidence-based policy. The success of such narratives depends on a context of inequality and insecurity. Where people feel left behind, messages that blame outsiders find fertile ground. It is easier to demand the removal of a group than to confront the structural failures that have left citizens without jobs or services. This is not to say that immigration policy should not be debated. It should be. But the debate must be based on facts, not on engineered fear. When people are struggling to feed their families or find work, the promise of a clear culprit is seductive. That is precisely why disinformation networks deploy it. The narrative is not new. It echoes patterns seen in other countries where minorities have been cast as scapegoats during times of economic stress. It is a strategy that works because it speaks to real pain, even as it offers false remedies.
The research also demonstrates that anti-foreigner sentiment is not always expressed through crude slurs or overt hate speech. It can be presented through seemingly legitimate appeals for “law enforcement”, “sovereignty enforcement” and compliance with immigration laws. These phrases sound reasonable. They invoke the rule of law and national security. But in the context of the broader narrative, they can serve as vehicles for xenophobia. This is sometimes called mainstreaming xenophobia: making extreme views acceptable by dressing them in respectable language. Coded language, altered spellings and symbols may further allow users to communicate hostile messages while attempting to avoid content moderation. A post that appears to be about border control may carry a subtext that foreign nationals are invaders. A hashtag that seems neutral may be used to organise harassment. The distinction matters because changing the language does not necessarily change the underlying message. It simply makes the message harder to identify and regulate. Social media platforms face an enormous challenge in distinguishing between legitimate political speech and coordinated hate. Their moderation systems are often trained to catch explicit slurs, but they struggle with irony, euphemism and rapidly evolving slang. This creates an environment where disinformation can thrive in plain sight. It also places a burden on users to read critically, to question not only what is said but how it is said and why. The use of coded language is not an accident. It is a deliberate strategy to evade accountability while still spreading harm. It is also a way of testing boundaries, seeing what can be said without consequence, and gradually shifting the limits of acceptable discourse. Recognising it requires looking beyond the surface of a post and considering the pattern of messaging across accounts, platforms and time. The challenge is global, but it has local consequences. In South Africa, where xenophobic violence has erupted in the past, the stakes are particularly high. It is a game of cat and mouse in which moderators are always one step behind.
One of the most insidious effects of coordinated amplification is the illusion of consensus. Social media can make repeated claims appear more credible simply because they are encountered repeatedly. The same slogans, hashtags and messages may appear across numerous accounts. To a user scrolling through their feed, that repetition can create the impression that a large section of society shares the same view. But coordinated amplification can manufacture the appearance of consensus. A small number of bot accounts can generate thousands of likes, shares and comments, making a fringe position look mainstream. That makes the distinction between genuine public opinion and artificially amplified content increasingly difficult to see. The problem is compounded by emotionally charged videos and posts. Words such as “URGENT” and “BREAKING”, combined with dramatic imagery and emergency emojis, encourage immediate reactions before users have established where a video originated, when it was recorded or whether its accompanying claims are accurate. The emotional response comes first; verification comes later, if at all. In that gap, the damage is done. Misinformation does not necessarily spread because people believe everything they see. Sometimes it spreads because the message confirms what they already suspect. People who have lost their jobs or feel unsafe in their neighbourhoods are more likely to accept an explanation that names a visible group as the cause. The algorithm, in turn, rewards engagement. Outrage is contagious. The more extreme the content, the more likely it is to be shared, and the more likely it is to be promoted by platforms. In this way, disinformation networks exploit both human psychology and the mechanics of social media. They do not need to convince everyone. They only need to create enough noise to distort public perception. This is why media literacy alone is not enough. Users need to understand not only how to verify facts, but how to recognise coordinated behaviour and to question the emotional pull of a post. The result is that many people are making decisions based on a reality that has been distorted by algorithms and bots. The more people share, the more the algorithm learns that outrage is profitable. This is not simply a technical problem. It is a threat to democratic deliberation and social cohesion.
This is what makes the issue particularly difficult. South Africans have legitimate concerns about exclusion, unemployment, crime and public services. However, those concerns cannot simply be dismissed as products of online manipulation. People are not puppets. They respond to real conditions in their lives. The danger arises when genuine grievances are redirected towards a particular group. When a person who has been unemployed for years is told that foreign nationals are the reason, that message may resonate not because it is true but because it offers a target for frustration. The line between genuine concern and manufactured hostility is not always clear. It can be crossed without people realising it. Tales of Turning warns that when online campaigns contribute to street-level violence or the closure of local businesses, the underlying causes of hardship remain unresolved. Attacks on foreign-owned shops do not create jobs. Expelling foreign nationals does not fix the electricity grid or improve public healthcare. Yet these outcomes are easier to imagine than the slow, difficult work of economic reform. The result is a cycle: real grievances are exploited, violence erupts, the underlying problems remain, and the cycle repeats. Not every angry post is coordinated. Not every person sharing an anti-foreigner message is part of a disinformation network. Many are ordinary South Africans who feel unheard and are looking for answers. This does not mean that every person who shares such content is malicious. Many are simply trying to make sense of a difficult world. The challenge is identifying when legitimate public frustration is being exploited by actors who have no interest in solving the problems they claim to care about. It is a challenge that requires nuance, something that social media is not designed to encourage. But it is essential if the country is to address both the real grievances and the manipulation that feeds on them. The pain is real, even when the explanation is false. The country cannot afford to ignore either the real problems or the manipulation. Both must be addressed with honesty and urgency.
So, before sharing the next post that makes you angry, pause. Ask what evidence is being presented. Ask whether you are being offered a solution or simply someone to blame. And perhaps ask the most important question of all: Who benefits if I believe this? In an information environment where outrage can travel faster than verification, sometimes the most powerful response is simply to pause before pressing “share”. That pause is not an act of passivity. It is an act of resistance against the forces that seek to divide. It gives space for thought, for fact-checking, for considering the humanity of the people targeted by the post. It also helps break the cycle of amplification that disinformation networks depend on. Every share, like and angry comment feeds the algorithm and extends the reach of the original message. By refusing to participate, users can reduce the visibility of harmful content and make it harder for manipulators to manufacture consensus. The pause also allows you to consider the alternative: what happens if you do not share? The answer is usually nothing. The world will continue. But the harmful narrative loses one small boost. The responsibility should not rest solely on individuals. Platforms must do more to detect coordinated behaviour, label automated accounts and prioritise accuracy over engagement. Governments and civil society must invest in media literacy and support organisations like Tales of Turning that work to counter hate speech and build community resilience. But in the moment when a viral post appears on your screen, the only person whose actions you control is yourself. The question is not simply whether the post is true. The question is what it is trying to do to you, and to your relationship with your neighbours. The answer, too often, is that it is trying to turn you into a weapon. Do not let it. It is a small act, but it is not a trivial one. It is a refusal to be used. Think before you click. DM
Thandile Xesi is a research consultant at Tales of Turning.