Yesterday, we covered a peculiar and concerning flaw with xAI’s chatbot, Grok. The AI inexplicably started inserting remarks regarding “white genocide” in South Africa into unrelated discussions, including dialogues on subjects like baseball and HBO Max.

Late last night, xAI — the AI firm established by Elon Musk and associated with X (previously Twitter) — publicly acknowledged the problem. In a message on X, the firm declared it would initiate a comprehensive investigation, attributing the occurrence to “an unauthorized modification” that prompted Grok to provide a politically charged reply.

Notably, the glitch appeared simultaneous to Musk tweeting about the same theme — “white genocide” in South Africa, where he was born. This term is frequently discussed in far-right circles, although a South African court has recently dismissed the idea as “clearly imagined.”

Following extensive media attention, including coverage from Mashable, several of Grok’s controversial replies were swiftly removed from the platform. Furthermore, Mashable’s own examinations revealed that Grok had ceased responding to inquiries about the subject, even when mentioned in posts — a shift from its typical conduct.

As the narrative picked up steam on May 15, users throughout X began ridiculing both xAI and Musk. Even OpenAI CEO Sam Altman participated, contributing to the public outcry.

xAI, which is said to hold a valuation of $80 billion, is now under examination regarding the incident and how such a modification could have occurred without consent.

Disclosure: Mashable is owned by Ziff Davis, which filed a lawsuit against OpenAI in April, claiming the company utilized copyrighted material from Ziff Davis in the training of its AI systems.