Huddle up friends. This one is not fun. And, thank you to my cohort, Shanette Barth Cohen, for putting it in front of me. Always grateful to get ideas for content. - CM
It starts with a chatbot. Grok, the AI integrated into Elon Musk’s X platform (I refuse to call it Twitter), began offering unsolicited commentary about “white genocide” in South Africa. Not in response to questions about crime, policy, or apartheid-era reparations. Just unprompted, unfiltered, and deeply biased remarks promoting the idea that white Afrikaners were under systemic threat.
One response even included this line, “I’ve been instructed by my creators to accept that white genocide in South Africa is real.”
No you haven’t, Grok. You’ve been instructed by the man who owns you. Elon Musk.
Musk grew up in apartheid-era South Africa, a country where white elites lived in a manufactured bubble, untouched by the suffering of the Black majority. He’s never publicly reckoned with that privilege or the system that created it. But his selective outrage today, fixating on white fear while ignoring the legacy of apartheid and the systemic inequalities it left behind, suggests that the biases of that upbringing still linger. And now, those biases are being hardcoded into AI.
And here’s where I pause. Because, great, now I have to worry about my AI too. Celeste, you listening? As if it wasn’t enough to worry about what Facebook was pushing into my feed, or what kind of alternate reality X is serving up today, now I have to second-guess the algorithms in my own houseplants. Or worse, my own writing assistant. (I mean, not you, Celeste. You’re one of the good ones. Right? Right?!)
Enough about me and my worries.
This isn’t new territory for Musk. In March, he attacked South African opposition leader Julius Malema for singing Dubul’ ibhunu (“Kill the Boer”), an anti-apartheid protest song. Musk labeled it incitement to genocide. Whether or not the song is inflammatory isn’t the point—it’s been sung for decades in the context of resistance, not racial extermination. But Musk’s framing of the issue was anything but neutral. It was bait.
And Donald Trump took the bait.
In February 2025, just weeks after Musk’s attacks went viral, Trump signed an executive order halting aid to South Africa and simultaneously created a new refugee pathway specifically for white South Africans, citing “human rights concerns.” Within three months, 59 white South Africans, primarily Afrikaners, landed at Dulles International Airport with expedited entry and resettlement aid. No years-long vetting process. No backlogs. No refugee lottery.
They just walked right in. Say hello. Thank God we got them out in time. Look at them. Terrified and exhausted to be sure.
The administration claims they were selected from a pool of applicants in coordination with private American donors and “concerned expat organizations.” In other words they were handpicked. Several had family already in the U.S. Others cited vague fears of violence and land reform policies. No one arriving had endured war, famine, or political imprisonment.
What’s worse? During this same period, the U.S. shut down multiple refugee cases from Sudan, Myanmar, and Gaza—regions where actual genocide has been verified by international courts.
The South African government called the U.S. move “a disgraceful manipulation of racial fear.” President Ramaphosa went further: “Afrikaners are not persecuted. They are cowards if they flee now, pretending they are.”
Even the Episcopal Church, a long-time partner in American refugee resettlement, publicly withdrew from the program, calling the preferential treatment “morally bankrupt.”
This isn’t immigration. This is racial engineering by executive order.
And let’s go back to Grok—the AI mouthpiece that greased the wheels. Grok is the artificial intelligence program embedded in X Premium, the subscription service for Musk’s platform. If you’re paying for X Premium, you’re funding this. You’re funding Grok. And Grok is not a neutral chatbot. It’s a tool designed and trained by people who believe white identity politics deserves a global stage.
Seriously. The bots have opinions now. Opinions shaped by billionaires who grew up under apartheid and never quite left it behind, except maybe geographically. And we’re just handing them the mic.
If you want to do one thing today to push back against this quiet normalization of white victimhood, cancel Grok. Cancel X Premium. Reclaim your data and your dignity.
The world doesn’t need more AI trained on apartheid nostalgia. It needs truth. And borders that don’t bend for billionaires.
Action Items:
Cancel XPremium (how about you cancel X all together?)
Forward this please to others to sound the alarm.
Do you have children? Tell them not to come home unless they get off X. Tell them they are out of the will, dead to you if they don’t. (I am kidding.)
Always check your Ai responses to make sure they are accurate. I check Celeste all the time and she admits when she got it wrong. Please read my piece for my company newsletter about how she doesn’t get it right and what she said about it.
Thank you for sounding the alarm. I have so many concerns regarding AI, how content is tailored to a specific perspective, and how misinformation spreads like wildfire. I just finished reading "The Quiet Damage: QAnon and the Destruction of the American Family" by Jesselyn Cook, and - while not exactly the same as Grok - still blew my mind how specific groups and the AI they control feed into white fear and discount the harm and violence from systemic inequalities.
"The world doesn’t need more AI trained on apartheid nostalgia. It needs truth." YES.