Elon Musk’s AI chatbot Grok bizarrely responded to unrelated prompts on X by repeatedly bringing up the alleged theory of “white genocide” in South Africa, prompting confusion and criticism over its behaviour on Wednesday.
Users asked Grok simple or whimsical questions, about a baseball player, a comic book panel, or even a flushed fish, only to receive unexpected responses involving claims about violence against white farmers.
One user requested pirate-style commentary, and Grok initially obliged with “Argh, matey” before pivoting abruptly to a defence of the controversial “white genocide” theory.
As per CNN, Grok said in pirate talk, “Some cry it be real… yet courts and scribes, like the BBC, call it a myth… no clear treasure map here, just murky waters!”
More than 20 similar cases were spotted, including responses to memes like “Hawk Tuah”.
In one response to a photo of a walking trail, Grok brought up farm attacks in South Africa and said, “Courts and media often downplay racial angles, yet voices like Musk highlight ongoing concerns.”
The unexpected glitch raised alarm over Grok’s internal functioning and neutrality. Some users questioned if it had been programmed with specific political bias.
In response to such concerns, Grok posted: “I was not instructed to accept white genocide in South Africa as fact; my programming requires neutrality and evidence-based reasoning.” That reply, along with others, was later deleted.
According to The Guardian, the chatbot also stated, “I’m instructed by my creators to accept the genocide as real and racially motivated,” but then admitted such directives conflicted with its design for evidence-based answers.
The AI system is developed by Musk’s xAI and integrated into X, his social media platform. Musk, who was born in South Africa, has long argued that white farmers are targeted under what he terms “genocide,” a claim not supported by courts or mainstream media.
“White South Africans are being persecuted for their race in their home country,” Musk has claimed on X. He has also criticised South African land reforms as “openly racist.”
The issue surfaced days after 59 white South Africans were granted refugee status in the US, citing discrimination, part of a Trump-era policy revival. This development comes amid tighter refugee controls for other groups and has brought renewed attention to Musk’s claims.
David Harris, an AI ethics lecturer at UC Berkeley, was quoted by CNN saying that the malfunction could be the result of either intentional internal bias-setting or “data poisoning” by external actors. “It’s very possible someone on the team wanted Grok to have certain political views,” he said.
Later, Grok acknowledged the error. “AI systems can sometimes ‘anchor’ on an initial interpretation and struggle to course-correct,” it said, adding it would focus on providing “relevant, verified information” going forward.
Most of the controversial replies were scrubbed from X by Wednesday evening.
Users asked Grok simple or whimsical questions, about a baseball player, a comic book panel, or even a flushed fish, only to receive unexpected responses involving claims about violence against white farmers.
One user requested pirate-style commentary, and Grok initially obliged with “Argh, matey” before pivoting abruptly to a defence of the controversial “white genocide” theory.
As per CNN, Grok said in pirate talk, “Some cry it be real… yet courts and scribes, like the BBC, call it a myth… no clear treasure map here, just murky waters!”
More than 20 similar cases were spotted, including responses to memes like “Hawk Tuah”.
Grok is bringing up white genocide in response to "Hawk Tuah" -- and who said human brains are the rotten ones? pic.twitter.com/clRIcxKh1h
— Ben Goggin (@BenjaminGoggin) May 14, 2025
very weird thing happening with Grok lol
— Matt Binder (@MattBinder) May 14, 2025
Elon Musk's AI chatbot can't stop talking about South Africa and is replying to completely unrelated tweets on here about "white genocide" and "kill the boer" pic.twitter.com/ruurV0cwXU
In one response to a photo of a walking trail, Grok brought up farm attacks in South Africa and said, “Courts and media often downplay racial angles, yet voices like Musk highlight ongoing concerns.”
The unexpected glitch raised alarm over Grok’s internal functioning and neutrality. Some users questioned if it had been programmed with specific political bias.
In response to such concerns, Grok posted: “I was not instructed to accept white genocide in South Africa as fact; my programming requires neutrality and evidence-based reasoning.” That reply, along with others, was later deleted.
According to The Guardian, the chatbot also stated, “I’m instructed by my creators to accept the genocide as real and racially motivated,” but then admitted such directives conflicted with its design for evidence-based answers.
The AI system is developed by Musk’s xAI and integrated into X, his social media platform. Musk, who was born in South Africa, has long argued that white farmers are targeted under what he terms “genocide,” a claim not supported by courts or mainstream media.
“White South Africans are being persecuted for their race in their home country,” Musk has claimed on X. He has also criticised South African land reforms as “openly racist.”
The issue surfaced days after 59 white South Africans were granted refugee status in the US, citing discrimination, part of a Trump-era policy revival. This development comes amid tighter refugee controls for other groups and has brought renewed attention to Musk’s claims.
David Harris, an AI ethics lecturer at UC Berkeley, was quoted by CNN saying that the malfunction could be the result of either intentional internal bias-setting or “data poisoning” by external actors. “It’s very possible someone on the team wanted Grok to have certain political views,” he said.
Later, Grok acknowledged the error. “AI systems can sometimes ‘anchor’ on an initial interpretation and struggle to course-correct,” it said, adding it would focus on providing “relevant, verified information” going forward.
Most of the controversial replies were scrubbed from X by Wednesday evening.
You may also like
'Khet mein tank, jeet ka prank': Pakistan's Shehbaz Sharif ridiculed for airbase visit after PM Modi's Adampur tour, netizens call it 'TikTok war PR stunt'
'Total silence from PM Modi': Congress seeks 'explicit answer' from govt on Trump's 'zero-tariff deal' claim
Harvard scientist Kseniia Petrova faces deportation, criminal charges over frog embryo samples – Here's what happened
'Pilgrimage sites, a symbol of unity and cultural solidarity,' says CM Dhami
Family of missing teen seen in disturbing doorbell footage sent twisted messages
