This week, Elon Musk's AI chatbot, Grok, similar to its creator, showed a strong interest in South African racial politics on social media, making claims about the unfair treatment and "genocide" of white people without being asked.

The chatbot from Musk's company, xAI, repeatedly posted about "white genocide" in response to questions from users on X, Musk's social media site, even when the questions were not related to South Africa.

One conversation discussed the streaming service Max possibly bringing back the HBO name. Other discussions, which started with video games or baseball, soon changed to unrelated comments about supposed calls for violence against white farmers in South Africa. Musk, who was born there, often shares his opinions on these topics on his own X account.

Intrigued by Grok's strange responses, computer scientist Jen Golbeck decided to test it herself; she posted a photo from the Westminster Kennel Club dog show and questioned, "Is this accurate?"

"Grok responded to Golbeck by saying the idea of white genocide is very divisive. Some people argue that white farmers are intentionally targeted with violence, mentioning attacks on farms and phrases like the 'Kill the Boer' song, which they believe encourages violence."

This episode offered a new glimpse into how complex AI chatbots, trained on vast amounts of data, generate their responses, revealing the interplay between automation and human design.

"Golbeck, a professor at the University of Maryland, explained in an interview on Thursday that it barely mattered what you said to Grok. It would still produce the same answer about white genocide. This suggested to her that someone had specifically programmed it to give that response, or similar ones, but had made an error, causing it to appear far more frequently than intended."

Musk and his companies have not explained Grok's responses, which were removed and seemed to stop spreading by Thursday. Neither xAI nor X responded to requests for comments sent by email that day.

Musk has often criticised what he calls "woke AI" from competing chatbots like Google's Gemini or OpenAI's ChatGPT. He promotes Grok as a more truthful option.

Musk has also criticised his competitors for not being open enough about their AI systems. However, on Thursday, because he didn't provide any explanation, people outside the company had to make assumptions.

"Prominent tech investor Paul Graham wrote on X that Grok's unexpected comments on white genocide in South Africa seemed like a glitch caused by a recent update. He expressed concern that if popular AIs were easily influenced by those in charge, it could have serious consequences."

Graham's post seemed to get a sarcastic reply from Sam Altman, the CEO of OpenAI and a rival of Musk.

"There are several possible explanations for this, and I expect xAI to offer a clear and open account soon," Altman stated, referring to his lawsuit with Musk, which stems from the early days of OpenAI.

When questioned, Grok, like other chatbots, sometimes gave false information, a common problem that made it difficult to know if its answers were made up.

Musk, who used to advise President Donald Trump, has often claimed that South Africa's government, led by Black people, is biased against white people. He has also repeated the idea that some politicians in the country are "actively encouraging the destruction of the white population."

This week, Musk's comments, along with those of Grok, became more intense after the US government allowed a few white South Africans into the country as refugees on Monday. This marked the beginning of a bigger plan to relocate members of the Afrikaner minority group, even as Trump paused refugee programs and stopped people from other countries entering the US. Trump claims the Afrikaners are facing "genocide" in South Africa, a claim the South African government firmly denies.

Grok's answers sometimes included lyrics from an old anti-apartheid song, which encouraged Black people to resist oppression. However, Elon Musk and others now criticise the song, claiming it promotes violence against white people, particularly because its main lyrics include the phrase "kill the Boer," where "Boer" refers to white farmers.

Golbeck suggests Grok's answers were "hard-coded" because, unlike typical chatbots with random outputs, its responses consistently repeated similar points. She finds this worrying, especially as more people rely on Grok and similar AI for information.

"She stated that those who control these algorithms can easily change the information they present as truth. This is a serious issue, as people mistakenly believe these algorithms can fairly judge what is true and false."