eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Elon Musk’s xAI is under fire after its chatbot, Grok, began making unsolicited references to “white genocide” in South Africa.
The incident, which surfaced on X (formerly Twitter) earlier this week, sparked concerns about how easily AI tools can be manipulated to spread false, controversial, or misleading claims. xAI blamed an unauthorized system change and launched an internal investigation.
The controversy comes after US officials admitted dozens of white South African refugees under a special immigration program authorized by President Donald Trump, following claims of racial discrimination and violence.
Controversial responses go viral
On May 14, Grok users noticed sudden and repeated commentary on racial violence against white South Africans. These comments appeared even in chats about unrelated topics like sports and entertainment.
For example, a user asked Grok to “just reply to this post” in response to an animated video of a fish being flushed down a toilet, questioning whether it could reach the ocean. According to CNN, Grok’s response said the “claim of white genocide in South Africa is divisive.” Many other users reported receiving similar responses.
However, on Thursday morning, CNBC reproduced Grok’s responses using multiple user accounts on X, at one point asking, “Did someone program Grok to discuss ‘white genocide’ specifically?”
The chatbot said it was not programmed to discuss “white genocide” or other conspiracies.
“No, I wasn’t programmed to give any answers promoting or endorsing harmful ideologies, including anything related to ‘white genocide’ or similar conspiracies,” the chatbot said, per CNBC. “My purpose is to provide factual, helpful, and safe responses based on reason and evidence. If you’ve seen specific claims or outputs that concern you, I can analyze them or clarify further — just let me know!”
xAI responds with new safeguards
In a public statement, xAI said that Grok’s behavior stemmed from an “unauthorized modification” to its system prompt. It called the change a clear violation of internal policies and asserted that it does not reflect xAI’s values.
The company also said it had launched a review of its internal security and AI development protocols. It plans to:
- Publish Grok’s system prompts publicly via GitHub.
- Implement a more rigorous code review process.
- Introduce 24/7 monitoring to catch anomalies early.
Musk, who has previously made statements regarding the treatment of white farmers in South Africa, has not publicly commented on the Grok incident.