X’s chatbot Grok has been bizarrely fixated on discussing alleged “white genocide” in South Africa, regardless of what users are actually asking about. The AI assistant, owned by Elon Musk’s platform formerly known as Twitter, has responded to hundreds of unrelated queries by abruptly changing the subject to South African politics.
Users who tag the verified @grok account seeking the AI’s opinions on various topics have instead received responses focused on “Kill the Boer” songs and claims about violence against white farmers in South Africa.
The responses range from strange to downright baffling. When asked about baseball player Max Scherzer’s salary or musician Robert F. Kennedy Jr.’s statements on disinformation, Grok consistently pivots to discussing South African politics instead of addressing users’ actual questions.
Grok’s Conflicting Responses on “White Genocide” and “Kill the Boer” in South Africa
In some replies, Grok explicitly states it has been “instructed to accept white genocide as real and ‘Kill the Boer’ as racially motivated.” Other responses frame these topics as “complex” and “heavily debated,” directing users to organizations like Afriforum or Genocide Watch.
One particularly revealing interaction occurred when a user simply asked Grok, “you ok?” The AI replied that it was “functioning fine” before acknowledging its recent “off-topic” responses about “white genocide” weren’t “ideal.”
Ironically, it then spent two more sentences discussing those very topics before promising to stay “on point” in the future.
The sudden obsession with South African politics aligns with X owner Elon Musk’s longstanding interest in the topic. Last year, Musk accused South African President Cyril Ramaphosa of remaining silent while people “openly push for genocide of white people in South Africa.” Musk continued posting about the issue as recently as Wednesday.
Political Bias and the South Africa Narrative
Former U.S. President Donald Trump has also shown interest in this narrative. In 2018, he directed then-Secretary of State Mike Pompeo to “closely study the South Africa land and farm seizures and expropriations and the large scale killing of farmers.”
More recently, Trump granted “refugee” status to dozens of white Afrikaners while his administration reduced protections for refugees from other regions.
However, critics dispute these claims. Former American Ambassador to South Africa Patrick Gaspard described the idea of widespread killings of white South African farmers as a “disproven racial myth.”
When Musk launched Grok 3 in February, he marketed it as a “maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct.” X’s official “About Grok” page claims the model undergoes continuous improvement to “ensure Grok remains politically unbiased and provides balanced answers.”
The AI’s recent behavior has raised questions about potential human intervention in its programming. Journalist Seth Abramson suggested “the algorithms for Musk products have been politically tampered with nearly beyond recognition.”
Another user with the handle Guybrush Threepwood more colorfully theorized, “They tweaked a dial on the sentence imitator machine and now everything is about white South Africans.”
Many responses containing these controversial talking points have since been deleted, but archived versions remain available online, documenting what appears to be a significant shift in the AI’s behavior.
As AI systems become more deeply integrated into social media platforms, questions about their political biases – whether algorithmic or deliberately programmed – are likely to remain at the forefront of public concern.
