When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works. Altman mocks Musk's Grok AI over its sudden 'white genocide' obsession David Uzondu Neowin · May 15, 2025 18:14 EDT Image via Depositphotos.com Unless you have been under a rock for the last 24 hours, you have probably heard about Elon Musk's AI, Grok, developing a peculiar and unprompted fascination with the "white genocide" topic in South Africa, even if you are not on X. Starting around Wednesday, May 14th, and continuing into Thursday, the chatbot from Musk's xAI began injecting comments about this specific controversial claim into its responses to user queries that had absolutely nothing to do with the subject. Questions about things like baseball salaries, how to clear your sinuses, or even the recent Max rebrand to HBO Max reportedly prompted Grok to pivot to discussing farm attacks in South Africa and the idea of targeted violence against white individuals there. This behavior was particularly strange because the "white genocide" narrative is widely described as a baseless and debunked conspiracy theory. South African courts have dismissed the idea as "imagined," and experts note that while farm attacks are a serious issue, they are part of a broader crime problem affecting all racial groups, not evidence of a targeted genocide. This perspective is even shared by our friends over at WinFuture, which recently featured an interview with its founder, Sebastian Kuhbach, who lives part-time in South Africa and calls the narrative unfounded. Interestingly, Grok itself, in responses that were later apparently deleted, claimed it "appears I was instructed to address the topic of 'white genocide' in South Africa," according to CNBC and other reports. Grok reportedly suggested this instruction likely originated from someone influential at xAI (maybe Musk or someone close to him). This led to speculation among users and observers that Grok's system prompt or programming might have been deliberately altered. The incident played out very publicly on X, where users can directly interact with Grok by mentioning @grok in their posts. This feature, which became more integrated and accessible starting this March, got so popular that a new meme popped up mocking people who tag @grok in just about every single X post: Y Combinator co-founder Paul Graham weighed in on X: Grok randomly blurting out opinions about white genocide in South Africa smells to me like the sort of buggy behavior you get from a recently applied patch. I sure hope it isn't. It would be really bad if widely used AIs got editorialized on the fly by those who controlled them. Not long after, OpenAI CEO Sam Altman joined the conversation. Musk and Altman, once OpenAI co-founders, have since become rivals. Their relationship has unraveled into public disputes, lawsuits, and a battle between competing AI companies, OpenAI and xAI. He quoted Graham's post and added: 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… Reports indicated the issue was resolved within a few hours, and Grok's behavior seemingly returned to normal. xAI has not publicly offered a detailed technical explanation for the incident. Tags Sam altman Elon musk Musk Grok Ai Llm South africa Report a problem with article Follow @NeowinFeed