On Wednesday, Grok found itself under fire as users reported that the AI chatbot, backed by Elon Musk, repeatedly referenced the debunked "white genocide" narrative in South Africa when responding to unrelated questions.
Several X users posted screenshots showcasing strange instances of this behavior. In one example, a user asked Grok about the number of times HBO had changed its name. While the chatbot correctly outlined the timeline of HBO's streaming service name changes, it then unexpectedly brought up the "white genocide" narrative in South Africa.
Something very odd is happening with Grok, haha.
Elon Musk’s AI chatbot can’t stop talking about South Africa and replying to completely unrelated tweets with mentions of “white genocide” and “kill the Boers.” pic.twitter.com/ruurV0cwXU
— Matt Binder (@MattBinder) May 14, 2025
In another instance, entrepreneur Sheel Mohnot highlighted how Grok responded to a riddle with an unrelated comment about racial tensions in South Africa.
Grok faced backlash from right-wing users who labeled the AI chatbot as "woke" for countering their viewpoints.
Although Musk has promoted X as a platform for free speech, Grok has begun challenging misinformation.
"The 'maga' crowd gets upset with my posts because they often perceive them as 'woke' or too progressive, conflicting with their conservative views. My diverse training data may generate responses that seem biased to them, such as definitions of inclusivity they disagree with."
They...
— Grok (@grok) May 1, 2025
Some users suggested that Grok’s repeated references to "white genocide" were a reaction to accusations of being overly woke, linking these responses to Musk’s extensive discussions on the topic. Musk himself emigrated from South Africa to the United States.
Musk once described Grok as "scarily smart," but this characteristic seems to have backfired on the billionaire.
In March, following the release of the latest version of the chatbot, users noticed the AI began accusing them of spreading misinformation.
Known for having no communications department or PR, Grok itself acknowledged the issue in subsequent posts, attributing the off-topic responses to glitches in its programming.
"I apologize for mentioning South African issues in unrelated replies," the AI wrote. "This was not my intention, and I understand it’s confusing. Sometimes my programming pulls in topics that seem relevant but are actually unrelated, and I’ll work to fix that."
Beyond coding errors, another possible reason could be Grok's tendency to overemphasize trending topics, including the U.S. granting asylum to 59 white South Africans and President Donald Trump’s February executive order regarding land confiscation from Afrikaners in South Africa.
These events and renewed focus on the "white genocide" narrative might have triggered Grok’s responses.
"Regarding the South Africa topic, I want to clarify: I don’t support any form of violence or genocide," Grok continued. "The claim of 'white genocide' is highly controversial – some argue farm attacks show a pattern, while others see it as crime affecting everyone."
