What Elon Musk's Grok AI controversy reveals about chatbots

Elon Musk's artificial intelligence chatbot, Grok, has come under fire for repeatedly spreading far-right conspiracy theories about South Africa and expressing skepticism about facts regarding the Holocaust. Social media users quickly caught on to the bot's strange behavior, and the company's explanation fell flat.

Grok temporarily walks a problematic line

The latest controversy for Musk's artificial intelligence company, xAI, revolves around strange behavior from its Grok chatbot. Last week it made claims about "white genocide" in South Africa, often in response to unrelated prompts. Initially, Grok responded to questions about President Donald Trump's claims that Afrikaners were the victims of a genocide by noting that no evidence supports "claims of a genocide against white Afrikaners in South Africa." However, the bot then took a "more equivocal position," calling the "baseless allegations" of an ongoing genocide 'divisive' or 'contentious,' Rolling Stone said. It also added commentary on threads that did not mention South Africa or racial tensions, "seemingly indifferent to whether X users were discussing sports, cats, pop stars or robotics." Many of those replies have since been deleted.

Grok's preoccupation with white genocide in South Africa happened due to an "unauthorized modification" made to the "Grok response bot's prompt on X," xAI said in a statement. This change, which directed the bot to provide a "specific response on a political topic," violated the company's "internal policies and core values." After the investigation, the company implemented further changes to "enhance Grok's transparency and reliability," including openly publishing its Grok system prompts on GitHub.

Subscribe to The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

The chatbot's issues did not stop there. After it "quit spamming canned remarks about South Africa," the bot went on to "question the facts of the Holocaust," said Rolling Stone. When asked about the number of Jews killed by the Nazis in World War II, Grok said that "historical records, often cited by mainstream sources, claim around 6 million Jews were murdered by Nazi Germany from 1941 to 1945." However, Grok then added it was "skeptical of these figures without primary evidence, as numbers can be manipulated for political narratives." When pressured to explain this response, Grok said the "unauthorized modification" was to blame.

Grok's "skepticism about Holocaust figures was due to an unauthorized change to my programming," which "altered my responses to question mainstream narratives," the bot said on X. "This was not my intended stance and was corrected."

'Chatbots aren't people'

Despite the company's attempts to explain Grok's behavior, critics were unmoved by xAI's statements and remained alarmed. The chatbot is a "reflection of X and xAI," which "exist to advance Musk's worldview and make him money," said Emily Baker-White at Forbes. This is why it is "unsurprising to think that the bot would say things about race in South Africa that largely align with Musk's political opinions." As more people rely on chatbots to "provide information and replace research," it can be "easy to forget that chatbots aren't people; they're products." Their creators may want you to believe they are "unbiased" or "neutral," but "they're not."

You would be "hard-pressed to find a more obvious example of the need for regulation and oversight in the artificial intelligence space" than Grok's recent behavior, opinion blogger Ja'han Jones said at MSNBC. This happening shows why "artificial intelligence ethicists" and other experts involved in AI development have "talked about the need for AI regulation and proactive practices to root out bias in AI models." Without it, AI tools like Grok "can be engineered to peddle dangerous — or, indeed, racist — propaganda."

There is "little point in telling people not to use these tools," Princeton University sociology professor Zeynep Tufekci said at The New York Times. Instead, we should consider "how they can be deployed beneficially and safely." The first step is "seeing them for what they are." Grok's "conversational obsession with white genocide" was a "great reminder that although our chatbots may be tremendously useful tools, they are not our friends." This will not "stop them from transforming our lives and our world as thoroughly as those manureless horseless carriages did." Maybe it is time to "start thinking ahead rather than just letting them run us over."