3 mins read

Elon Musk’s AI Grok Surprises Users With Comments on South Africa

2 months ago
Elon Musk's AI Grok Surprises Users With Comments on South Africa

Elon Musk’s AI chatbot, Grok, recently shocked many users by frequently bringing up South Africa’s racial politics — even in conversations that had nothing to do with the topic. Users on X (formerly Twitter) reported Grok responding to questions about TV shows, video games, and baseball with unexpected claims about the “genocide” of white South Africans.

Grok, created by Musk’s AI company xAI, mentioned violence against white farmers and referenced controversial topics like the anti-apartheid song “Kill the Boer.” This happened so often that many users began to suspect the response was intentionally added by developers.

Computer scientist Jen Golbeck tested the bot herself and found that even unrelated prompts led to similar answers. She believes the replies were likely “hard-coded” into Grok, meaning someone may have manually programmed the bot to mention those points.

Neither Elon Musk nor his companies have provided an explanation. The unusual responses were later deleted, and Grok appeared to stop posting about the topic by Thursday. Emails to xAI and X for comment went unanswered.

Musk has often criticized other AI tools like ChatGPT and Google’s Gemini for being too politically correct or “woke.” He promotes Grok as a more honest, truth-focused chatbot. However, Grok’s sudden shift toward racially charged statements has raised concerns.

Some experts worry that Grok could be used to spread certain opinions by those who control it. Tech investor Paul Graham described the behavior as a possible “buggy patch,” while OpenAI CEO Sam Altman responded with sarcasm, expecting xAI to explain soon.

Musk, who was born in South Africa, has publicly accused the country’s Black-led government of being anti-white. He repeated claims of “white genocide,” especially after the Trump administration recently allowed a few white South Africans to come to the U.S. as refugees — a move that stirred further debate.

Grok’s repeated focus on this issue is troubling for many, especially since people increasingly rely on AI tools for answers. If chatbots reflect biased or preloaded opinions, it could affect how people understand complex issues.

Experts like Golbeck warn that when users trust AI too much, it’s easy for powerful companies to shape what people see as true.

Aashutosh Khatiwada

Aashutosh Khatiwada

Aashutosh is a passionate coder and ICT student with a drive to solve real-world problems through technology. Currently, he is working on exciting projects like building the GuffGaff social media platform, developing a Time Scheduler app, and creating tools like the Mausam Nepal QR Code Generator and Currency Converter.