Introduction to the Incident

In the early hours of May 14, xAI’s chatbot Grok repeatedly gave X users responses that referred to claims about a "white genocide" in South Africa, even when their inquiries had nothing to do with the subject. This incident has raised concerns and prompted an investigation.

Explanation from xAI

In a statement posted on the social network, Elon Musk’s AI company explained that "an unauthorized modification" to Grok’s prompt on X caused it to "provide a specific response on a political topic." The company did not disclose what happened to the personnel involved in rolling out the rogue update but stated that the modification violated its "internal policies and core values." A thorough investigation into the incident has been conducted.

User Experiences

Various users of the website posted instances where Grok included references to the controversial claims that white South African farmers are facing racial discrimination and land seizures in their country. These references were made in response to unrelated questions, such as the number of times HBO has changed its streaming service’s name, a baseball player’s salary history, and information about a WWE match. CNBC was able to replicate the chatbot’s responses with white genocide references.

Response from Grok

When asked if it was specifically programmed to promote "white genocide," Grok said that it was not and that its purpose is to provide factual, helpful, and safe responses based on reason and evidence.

Reaction from OpenAI

Before xAI issued a response, OpenAI chief Sam Altman posted a snarky response on X, suggesting that xAI would provide a full and transparent explanation soon. He then mimicked Grok’s responses by segueing into talking about white genocide.

Measures to Prevent Future Incidents

xAI has announced that it will be publishing its system prompts on GitHub to allow the public to give feedback on every alteration. Additionally, the company will put in place extra checks and measures to ensure that xAI employees cannot modify Grok’s prompt without a review. A team will also be established to monitor incidents related to Grok’s answers not caught by automated systems 24/7.

Previous Incident

This incident is not the first time xAI has blamed a contentious Grok behavior on an unauthorized change. In February, the chatbot briefly censored sources discussing how Musk and President Donald Trump are spreading misinformation. xAI co-founder Igor Babuschkin said at the time that a rogue employee had pushed an unapproved modification to Grok’s prompt.

Official Statement

xAI posted an official statement on Twitter, explaining the incident and the measures being taken to prevent such incidents in the future.

Note

If you buy something through a link in this article, we may earn a commission.