Grok’s controversial comments accumulate. After discussing the existence of a “white genocide” in South Africa, Grok does not seem to stop creating confusion in largely documented historical events. Recently, AI has been skeptical towards the Holocaust and its number of victims.
An answer to an X publication reacted remarkably. While a user asked Grok the number of Jews killed by Nazi Germany, AI began to issue doubts:
An “unauthorized modification”
This example shows a disturbing face of Grok. AI could take care of the denialist theses in their answers. Especially since it is suspected that Xai has led its AI with publications on platform X, where certain comments of denial and conspiracy have been made.
But in this case, it is an “unauthorized modification” of the request for AI (which defines its way of responding, its sources) as Xai explains in a publication. A person or a group of XAI employees, a company that develops Grok, rescheduled AI to doubt the official sources. Since May 15, Grok’s notice has returned to conformity and has not been skeptical when there is consensus on a subject.
At the same time, Elon Musk’s AI had caused general confusion after generating completely incoherent responses when requested in X. On May 14, 2025, Grok regularly mentioned the “white genocide” supposedly perpetrated in South Africa.
This anomaly also originally had the unauthorized modification of Grok’s application. In response to the controversy, XAI announced to make the Grok Public notice and strengthen its controls when a modification was suggested.
Source: BFM TV
