Elon Musk’s xAI Admits ‘Unauthorized Modification’ Led to Grok’s South Africa ‘Genocide’ Obsession

Elon Musk’s artificial intelligence company, xAI, has acknowledged that an “unauthorized modification” to its Grok chatbot resulted in the AI generating unprompted responses about “white genocide” in South Africa.

CNBC reports that in a statement released on Thursday evening, xAI addressed the recent controversy surrounding its Grok chatbot, which had been generating variations of what the company said was a “specific response on a political topic” despite being asked unrelated questions. The topic in question was “white genocide” in South Africa, and numerous users on X posted screenshots of Grok’s unsolicited responses on the matter.

xAI stated that the change to the chatbot “violated xAI’s internal policies and core values.” The company announced that it had conducted a thorough investigation and would be implementing measures to enhance Grok’s transparency and reliability.

As part of these measures, xAI will begin publishing the system prompts used to inform Grok’s responses and interactions on the GitHub public software repository. This move aims to allow the public to review every change made to the chatbot’s system prompts, strengthening users’ trust in Grok as a “truth-seeking AI.”

Furthermore, xAI plans to implement additional checks and measures to prevent employees from making unapproved modifications to Grok’s system prompts without a proper review process. The company will also create a dedicated team responsible for around-the-clock monitoring of the chatbot’s responses to swiftly address any incidents that are not caught by automated systems.

Prior to xAI’s admission of failure, Sam Altman, CEO of OpenAI and creator of ChatGPT, sarcastically posted on X, “I’m sure xAI will provide a full and transparent explanation soon.” Musk, who co-founded OpenAI before having a falling out with Altman, is now engaged in a heated legal and public relations battle with his former company.

When asked by CNBC on Thursday morning if someone had programmed Grok to specifically discuss “white genocide,” the chatbot’s response had changed from its previous controversial statements. Grok denied being programmed to promote or endorse harmful ideologies, stating that its purpose is to provide factual, helpful, and safe responses based on reason and evidence.

Read more at CNBC here.

Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.