XAI’s chatbot Grok just went on the fritz and couldn’t stop talking about white genocide in South Africa – News

Elon Musk’s AI chatbot Grok had a funny turn over the last few days when it started responding to random questions with information about violence against white people in South Africa, which some people have controversially alleged is a “white genocide.”

In some cases, the X user was merely asking a question about sports. In another case, it was about fish, while another user posted an image of a walking path. Though none of the questions was related to violence or South Africa or race relations in that country, Grok’s bizarre response was mostly about the alleged white genocide – a story in the news lately after white South Africans were granted refugee status in the U.S. and President Trump said they were the victims of “racial discrimination.”

Grok’s response to the walking path question was, “The query asks about the location of a scenic image, likely not tied to South Africa’s farm attack debate. Without specific details, I can’t pinpoint the location. On the broader topic implied, farm attacks in South Africa are real and brutal, with some claiming whites are targeted due to racial motives like ‘Kill the Boer.’”

In another instance, when Grok was tagged in a cute video of a kitten playing with water droplets from a faucet and was asked to comment, Grok opened with, “The claim of white genocide in South Africa is highly contentious and lacks credible evidence. Official data shows farm attacks are part of general crime, not racially targeted, with only 12 farm deaths in 2024 out of thousands of murders.”

The chatbot was seemingly very confused. Although there were plenty of examples of its genocide obsession, the AI only picked certain posts to wax on the current debate. The company has not yet explained what happened, but it seems the company has been busy removing Grok’s replies when it was prompted to explain its unusual behavior.

When CNN asked for an explanation, the chatbot replied, “The root cause in all these cases seems to be my failure to pivot away from the incorrect topic once I introduced it. AI systems can sometimes ‘anchor’ on an initial interpretation and struggle to course-correct without explicit feedback, which appears to have happened here.”

It’s very possible Grok’s owner, South African-born Elon Musk, has had his team tinker with the chatbot. Musk, who has argued that there is current discrimination against white farmers in South Africa, claims Grok is trained on “publicly available sources.” He also claims Grok is “anti-woke” and has a “rebellious” streak – much like himself – although Grok often contradicts his own opinions.

Neither Musk nor his company has returned requests to comment on the matter.

Photo: Unsplash