Not long ago, Kenya’s President William Ruto ordered the closure of small businesses run by foreign nationals. The targets were hawkers: the street vendors, market traders, people selling basic goods from stalls and carts across Nairobi’s informal settlements. “We have not built investor confidence for hawkers to come to Kenya,” Ruto said. Within days, hundreds of Burundian nationals were queuing outside their embassy seeking travel documents, harassment was being reported, and the government was scrambling to clarify that it had been misunderstood.
Sheila Owigo Olang, a political economist, spoke about the actual data, explaining that there is little evidence that removing foreign traders would create more opportunities for Kenyans. The policy, she said, was populist and timed ahead of elections.
The Kenyan government had watched South Africa do this for a while. Now it was trying the same move.
How It Is Done
Social engineering, in the political sense, is not simply propaganda. It is a specific technique: redirect a population’s legitimate grievance toward a target that is easier to reach and less dangerous to attack than the actual source of the problem.
The grievance in South Africa is very real. Unemployment in South Africa sits between 32% and 40% depending on your mode of measurement. The legacy of apartheid, specifically the concentration of land and capital in the hands of a small minority, has never been meaningfully addressed. The promises of the post-1994 transition have not been kept for the majority of black South Africans. That frustration is legitimate and has been building for thirty years.
But foreign nationals selling vegetables and running spaza shops are not why unemployment is 40%. A Pew Research poll found that 61% of South Africans believed that immigrants were more responsible for crime than other groups. Politicians and public figures repeatedly linked foreign nationals to gun running, drug trafficking, and armed robbery. But the data says otherwise. South Africans make up 89.5% of the prison population in the country. Foreign nationals are 10.5 percent. Peer-reviewed research has found no empirical evidence that immigrants take jobs from South Africans, and some categories of migrant work have actually been shown to increase employment opportunities for South Africans. A study published in 2025 found that citizens who are dissatisfied with government performance are still welcoming to some immigrants, suggesting they are more frustrated with government underperformance than with the presence of foreigners.
The data does not support the narrative. But the narrative persists anyway, because it is useful. A population focused on the Zimbabwean running a shop in their neighborhood is a population not focused on the politicians who have failed to deliver land reforms, functional public services, or meaningful economic redistribution since 1994.
The Same Trick, But A Different Flags
In Kenya, Ruto is governing a country with a restless, politically aware youth population that forced him into significant concessions just two years ago. The 2024 Gen Z protests over finance bill taxation demonstrated that Kenyan youth are capable of organized resistance when a grievance is clearly named. The move against foreign traders is, at least in part, a redirection: give the anger somewhere safer to go before the next election.
The script is identical to South Africa’s. The targets are people with even less political recourse: not citizens, not voters, not people whose removal would require genuine policy courage.
This is what social engineering looks like at its most efficient. It does not require elaborate technology or sophisticated messaging. It requires a politician willing to point at the most vulnerable people in the room and say: them. That is your problem. And when someone asks for evidence, the feeling of the thing is usually enough to sustain it. People feel that foreigners are taking jobs. People feel that crime is linked to immigration. That feeling does not require data. It only requires repetition, and a leader willing to amplify it.
Further Away, Same Logic
The American midterm elections are fast approaching. On one side of the campaign, the argument is about jobs, education, and the wealth gap: substantive policy disagreements about how a country should be organized. On the other side, a significant portion of the messaging has focused on a candidate’s Arab background and the suggestion that their election would bring Sharia law to the country. Not their policy platform. But their background, their name, and the implication of their ancestry.
This is not new in American politics. It is a version of the same mechanism: take a legitimate anxiety about the future, attach it to a person rather than a system, and make it about who someone is rather than what the actual problem is.
The Gaza situation operates on a related distortion, though in reverse. A genocide is being documented in real time. International courts have used the word. Journalists, aid workers, and United Nations officials have described what is happening in terms that leave little room for interpretation. And yet, a significant infrastructure of messaging has been constructed around a single rhetorical move: that naming what is happening makes you an antisemite. Not engaging with the evidence, nor responding to the specific arguments. Just naming the person who is speaking as a kind of person whose words do not need to be heard.
This past week, while on Ed Sheeran’s tour, Macklemore showed footage of Gaza, and led the crowd in chants of “Free Palestine.” The billionaire owner of a stadium that was part of the tour subsequently coordinated with other venue owners to have Macklemore dropped from the tour by Ed. At the moment, all of Sheeran’s opening acts have withdrawn from the tour in protest. What Macklemore did was express solidarity with people whose deaths are documented in international court proceedings. The response was not an argument, but rather the removal of his platform. That is the antisemitism charge operating as an infrastructure: not a claim made in good faith that can be responded to, but a mechanism for ensuring that certain voices never reach the room.
Who Benefits From This
The useful question in all of these cases is not who is being blamed. It is who benefits from the blaming. In South Africa, the beneficiaries of xenophobia are the political and economic class who have managed to avoid accountability for thirty years of governance failure by keeping the conversation focused on foreign nationals. In Kenya, the immediate beneficiary is a president trying to redirect youth frustration before an election. In the American midterms, the beneficiary is a political strategy that has calculated it cannot win on policy and so it tries to win on identity fear. In the Gaza situation, the beneficiary of the antisemitism charge is a government that cannot defend its actions on their merits and so needs to criticize those actions as unspeakable.
The AI panic fits the same structure, though the mechanism is different. The story that artificial intelligence is coming for everyone’s jobs, told at the scale and speed it has been told, serves the interests of technology companies raising capital at inflated valuations, media platforms optimized for engagement through fear, and a political class that can point at technology and its advancement as the reason some people are struggling, rather than the policy decisions that created those conditions in the first place. The displacement is real in some sectors. The apocalyptic framing serves purposes beyond describing what is actually happening.
None of this means the underlying anxieties are illegitimate. South Africans are right to be angry about unemployment and inequality. Kenyans are right to want economic opportunity. Americans are right to worry about the direction of their country. Palestinians are right to expect that their lives be treated as worth something. Workers are right to ask what technological change means for their futures.
The engineering is not the anxiety. Rather, it is where the anxiety gets pointed.
To See It Clearly
Social engineering works best on populations that have been taught to feel rather than to ask. It works best when the question “who benefits?” is not a reflex. It works best when the media environment is optimized for emotional reaction rather than analytical response.
The antidote is not cynicism. It is not the conclusion that everything is manipulation and therefore nothing can be trusted. That is its own trap: a different kind of engineered helplessness that benefits the same people.
It is the habit of asking, every time a narrative arrives pointing your anger at someone, whether the people directing your attention are themselves above that line of sight. Whether the target you are being given is actually capable of causing the problem you are experiencing. Whether the conversation you are being invited into is the one that actually needs to happen.
The oldest trick in the book is convincing people that their real enemy is their neighbor. It has worked across centuries and continents because it is genuinely effective. The only thing that disrupts it is the refusal to look where you are being told to look, and the habit of turning around to see who is standing behind you, pointing.