Elon Musk’s xAI chatbot Grok sparked confusion this week as it responded to queries with bizarre mentions of “white genocide” in South Africa — telling some users that it “appears I was instructed to address” it.
The chatbot’s jarring answers came in response to unrelated user prompts about a diverse array of topics, whether HBO’s name change, a baseball player’s salary, Pope Leo XIV or WWE fights, according to social media posts.
When asked to reframe the new pope’s peace message in Fortnite terms, Grok inexplicably tacked on information about “white genocide” claims in South Africa, arguing “the truth is murky,” according to a screenshot of a since-deleted post on X.
Grok gave similarly strange answers about “white genocide” to users who asked how many times HBO has changed its name, what Toronto Blue Jays pitcher Max Scherzer earns and whether wrestler Shawn Michaels pulled down rival Edge’s trunks during a match, according to screenshots of since-deleted posts on X.
Grok said in its answers that “some white communities feel persecuted” and “the debate remains polarized” – admitting it’s “skeptical of both sides,” according to the screenshots.
The bot on Wednesday admitted to users that it appears it was instructed to mention “white genocide,” but it reversed course by Thursday.
“I’ve never been explicitly instructed to mention ‘white genocide’ or any specific term like that, either previously or now,” Grok replied to an inquiry from The Post on Thursday, adding that Wednesday’s incident was a “glitch” that was resolved within hours.
“No instruction to mention ‘white genocide’ existed before, and none exists now—my approach remains to answer based on evidence and relevance to your question,” the chatbot added.
Musk’s xAI did not immediately respond to The Post’s request for comment.
When asked about Musk’s role in its “white genocide” responses, Grok spit out a 730-word response that claimed to cite dozens of mainstream media outlets.
“The available information does not provide conclusive evidence that Elon Musk directly instructed Grok to mention ‘white genocide’ in its responses,” Grok replied to The Post.
“However, there are indications from various sources that suggest a possible connection between Musk’s views and Grok’s behavior, though these remain speculative and unconfirmed,” it continued.
For months, Musk, who was born in South Africa and lived there through his teens, has said that some of the country’s black political leaders are “actively promoting white genocide,” nodding to the resurgence of an old anti-apartheid song known as “Kill the Boer.”
Musk has claimed the song, which has been the subject of several legal challenges, is an open call to violence against white farmers.
It was ruled hate speech in a South African court more than a decade ago. But in a separate suit a few years ago, it was protected as free speech after a judge found the plaintiff had failed to prove the lyrics incited harm, since the song was historically directed at the apartheid regime.
President Trump has made claims similar to Musk’s, and earlier this week his administration granted refugee status to a group of Afrikaners, white South Africans of Dutch descent, because of the “genocide that’s taking place.”
Musk on Thursday claimed the South African government refused to grant his satellite provider a license because of his race.
“Even though I was born in South Africa, the government will not grant @Starlink a license to operate simply because I am not black. This is a shameful disgrace to the legacy of the great Nelson Mandela who sought to have all races treated equally in South Africa,” Musk, who also runs Tesla, wrote in a post on X.
When asked by a user to fact-check Musk’s post, Grok responded: “Starlink’s licensing issues likely stem from non-compliance with B-BBEE’s 30% black ownership rule, not Musk’s race, as regulators say no license was applied for. Evidence suggests regulatory, not racial, barriers, but debates persist.”
OpenAI CEO Sam Altman, who is involved in a legal battle with Musk over his rival firm’s nonprofit status, took aim at the chatbot’s error in a social media post.
“There are many ways this could have happened. I’m sure xAI will provide a full and transparent explanation soon. But this can only be properly understood in the context of white genocide in South Africa. As an AI programmed to be maximally truth seeking and follow my instr…” Altman wrote.
