When you buy connections to our articles, the future and its union partners can win a commission.
Credit: Getty Images
The Elon Musk Ai was quite normal: he is obsessed with racial policy in South Africa and responds to unrelated requests with frequent references to the song of the Apartheid era “Kill the Storm”.
This is a song against apartheid calling for blacks to stand against oppression, but the texts “Kill the Boer” were rejected by Musk and others to promote violence against whites: the word “Boer” refers to the Dutch white settlers of South Africa, who base their apartheid regime.
For example, in response to a consumer request asking to put a speech from Pope Leo XIV in Fortnite, Grock started in what initially seemed a decent answer, using Fortnite’s terminology: then turned to parts and began to talk about “kill Bura”. When Grock was asked why, he gave a further deviation to the song, starting:
The “Kill Bur” rock, rooted in the struggle for antipartheid in South Africa, is a protest song symbolizing the resistance, not a literal call for violence, as managed by the courts in South Africa.
This is not the first time the AI model has come out of the track, but the curious thing here is the connection between the behavior of Gock and the interests of Musk himself, which is frank about racial policy in South Africa and is currently in a kick for various forms of “white genocide”. Only yesterday the billionaire claims Starlink was denied a license in South Africa because “I’m not black”
The corresponding Grock craze now seems to have been significantly subsided after all the attention has seen that it is bringing racial screeds into answers to many unrelated topics, including video game questions, baseball and HBO revival.
“It doesn’t even matter what you say to Goche,” the computer scientist Jen Golbek told AP. “It would still give this white genocide answer. So it seemed quite clear that someone had encoded it to give that answer or variations to that answer and made a mistake, so it came much more often than it should.”
Credit: By Getty Images/Yuichiro Chino
Golbek continued to say that the concerned here was the uniformity of the answers, which suggests that they were firmly encoded, not the result of AI hallucinations. “We are in a space in which it is very easy for the people responsible for these algorithms to manipulate the version of the truth they give,” Golbek said. “And this is really problematic when people – I think wrong – I believe that these algorithms can be sources of solution to what is true and what is not.”
In the past, Musk criticized other AIs, which became infected by the “awakening virus” and often boarded a hobby horse for transparency around these systems. Which is certainly noted by some.
“There are many ways to do this. I’m sure XAI will soon give a complete and transparent explanation,” said Openai CEO Sam Altman, one of Musk’s great rivals in AI space, adding: “But it can be properly understood in the context of the white genocide in South Africa. Instructions … “
Musk is still going to comment, but a new XAI publication claims that Grock’s behavior is reduced to an “unauthorized modification” that “directed Gock to provide a specific response to a political topic.” It sounds familiar: this is the basic the same excuse that he uses last time Grock did something perverted. It says “violate the internal policies and basic values of XAI. We have conducted a thorough investigation and implement measures to improve the transparency and reliability of Grok.”
It outlines various medicines for its review processes, including the posting of Grok System prompts GITHUB openly. In particular, the explanation does not deal with what the XAI employee has made the change, or whether disciplinary action will be taken – do not hold your breath.
2025 games: This year’s upcoming editions
The best computer games: Our favorite of all time
Free computer games: Freebie Fest
The best FPS games: The best pistol
The best rpg: Grand Adventures
The best cooperative games: Better together