Every day, millions of South Africans open X, formerly Twitter, to see what is happening in their country. They scroll past jokes, sports highlights, and political arguments, but increasingly they also scroll past a steady stream of anger aimed at foreign nationals. That anger feels raw and real, especially in a country where unemployment is crushing, basic services are unreliable, and so many people are simply trying to get by. But a new investigation by the non-profit organisation Tales of Turning suggests that much of that online fury is not as organic as it seems. After studying 377,167 posts collected between 1 August and 25 September, researchers found a clear and disturbing pattern of digital manipulation, one that is deliberately turning genuine grievances into targeted anti-immigrant hatred ahead of the 4 November local government elections. The report paints a picture of a country where fake accounts, automated bots, and overseas operators are quietly pulling the strings of public conversation. When protests broke out in Durban and Soweto, with cars set alight and foreign-owned shops looted, many of us watched in horror and asked: how did we get here? The report suggests that a large part of the answer lies not in the streets, but in a carefully engineered online ecosystem designed to keep communities on edge, frightened, and divided.
To understand the current wave of unrest, one has to go back to a landmark court ruling that very few people actually read but many people talked about. On 7 July, the Constitutional Court ruled in the case of Scalabrini Centre of Cape Town versus the Minister of Home Affairs. The court found that certain provisions of the Refugees Act were unconstitutional, specifically the parts that allowed asylum seekers to be excluded from the asylum process for technical reasons like entering the country illegally or failing to register at a refugee reception office within a set time. The court did not say that everyone who wants to come to South Africa can simply walk in. It did not grant asylum, permanent residency, or citizenship to anyone. It simply said that every person’s claim must be properly assessed on its merits before being thrown out on a technicality. That is a reasonable and humane legal position. But when the Department of Home Affairs issued a directive to implement the ruling, refugee reception offices became overwhelmed. There were queues, confusion, and administrative chaos. On 9 October, the department formally withdrew the directive and said it would develop a new approach. By then, however, the damage was already done. Online, false claims spread like wildfire: that the courts had opened South Africa’s borders, that foreign nationals were being handed free documents, that the country was about to be flooded. The legal reality was lost in a storm of panic. Within days, that digital panic had spilled onto the streets. In Durban and Soweto, protests erupted, cars were torched, roads were blockaded, and shops were vandalised. It is a sobering reminder that a misleading post shared from a living room can end with a burning building.
The Tales of Turning report goes much deeper than describing the protests. Using a combination of human-trained machine algorithms and social network graphing, the researchers tracked user activity, emotional tone, and platform metrics across X. What they found is alarming: at least a quarter of the top 500 most active accounts spreading anti-foreign-national content are classified as inauthentic disrupters or automated bots. That means that while everyday South Africans make up about 65 percent of the overall voices in the conversation, a small but hyperactive group of fake or coordinated accounts is doing the heavy lifting to shape the narrative. These accounts do not just post once or twice. They push out massive volumes of content, constantly injecting hostility into feeds and making it seem as though there is a sweeping national movement of xenophobic anger. The single most active voice in the dataset posted or reshared derogatory content about foreign nationals 1,494 times in under eight weeks. That alone tells you something. No ordinary person with a job, a family, and a life has time to post that many times about any topic, let alone one subject. The account publicly claimed to be based in Johannesburg, but X’s own backend metrics revealed that it was actually operating from outside South Africa via an overseas app store connection. That discovery is deeply uncomfortable. It suggests that at least some of the anti-foreign sentiment pushing South Africans onto the streets is being manufactured by people who are not South African and do not share the consequences of the violence they are inciting.
The manipulation is not just about volume, but also about structure. The researchers mapped connections between local social media users and high-profile international influencers, revealing a network that moves cross-border talking points into local conversations. Some of these international accounts have huge followings and openly political agendas. Others are what the report calls hidden influencers, accounts with smaller audiences that sit quietly at strategic points in the network, gently steering conversations across multiple communities without ever setting off moderation alarms. These hidden influencers are perhaps the most dangerous because they are almost invisible. They do not shout; they suggest. They do not call for violence; they plant seeds of fear and mistrust. Over time, they normalise ideas that would have seemed extreme a decade ago. The report notes that outright hate speech made up less than 1 percent of the hundreds of thousands of posts examined, but that does not mean the content was harmless. The algorithms still flagged 1,500 pieces of extreme rhetoric that went far beyond normal political debate, including direct calls to terrorise and use violence against foreign nationals. Underneath all of this lies an emotional current, carefully cultivated anger designed to keep people in a state of anxiety and outrage. The manipulation works because it attaches itself to real pain. The fear of losing a job, the frustration with crime, the feeling that the government is not listening, all of these are genuine. When digital puppeteers tap into those feelings, ordinary people begin to repeat messages that are not their own, convinced they are finally speaking their truth.
Why would anyone go to such lengths to manufacture division in South Africa? The report argues that the answer is political, and the timing is crucial. With local government elections just around the corner, the strategy of divide and rule is being used to exploit existing fault lines in society. Instead of addressing the real grievances that people have, such as unemployment, housing, and failing public services, these campaigns turn communities against each other. The goal is to erode trust in the governing coalition, to create chaos, and to make people feel so insecure that they lose faith in all political leadership. When communities are busy hating each other, they are not uniting to demand better leadership. When people are distracted by the threat of an outsider, they stop asking why crime is so high, why the electricity keeps failing, and why the economy is not growing. The report uses the term “weaponised grievances” to describe this phenomenon. It is not that the grievances are fake; they are painfully real. The weaponisation makes them worse by aiming them at the wrong target. Fatigue, frustration, and fear are harvested, then fed through a machine of bots and coordinated accounts, and returned in the form of violent protest. It is a deeply cynical process. The authors of the report remind us that as we scroll through our feeds and watch the smoke rise from our streets, we should ask ourselves a simple but urgent question: when we react to the anger on our screens, are we truly trying to solve South Africa’s deep, painful problems, or are we dancing to a tune composed by someone hiding thousands of kilometres away?
There is no easy solution to this kind of digital manipulation, but awareness is the first line of defence. We need to be more suspicious of the loudest voices on our timelines, especially those who seem obsessed with dividing people into us and them. We need to ask who benefits when we post, share, and follow in anger. We need to demand better from social media platforms, which have allowed bots and foreign actors to operate with almost no accountability. But we also need to do something harder: we need to resist the urgent pull of outrage. The people who run these campaigns rely on our emotional reaction. Every time we share a post without verifying it, every time we let a stranger on the internet convince us that our neighbour is our enemy, we are doing their work for them. The report from Tales of Turning is more than a warning; it is an opening. If we know that a quarter of the top voices spreading hatred are fake, then we no longer have to be fooled by them. We can look at an anti-foreign rant and see it for what it is: a cheap attempt to buy our anger with a currency we can never spend. The real challenge is not defeating foreigners or closing borders. It is fixing a system that makes so many South Africans feel that they are in a fight for survival. That fight should be directed toward the policies and leaders that have failed them, not toward innocent people who are also looking for a better life. In the end, the battle we are engaged in is not a battle between South Africans and outsiders. It is a battle between truth and manipulation, empathy and fear, and it will be won or lost in the choices we make every time we pick up our phones.