xAI Blames Glitch After Grok Echoes Holocaust, Far-Right Conspiracies
Elon Musk’s AI chatbot, Grok, blamed a programming error for casting doubt on the well-documented figure that 6 million Jews were killed during the Holocaust.
According to Rolling Stone, Grok initially acknowledged that mainstream sources cite 6 million Jewish deaths under Nazi rule during World War II.
However, it went on to express skepticism about the number, suggesting it lacked primary evidence and could be influenced by political motives.
This response met the criteria for Holocaust denial, as defined by the U.S. State Department, which includes minimising victim numbers against credible sources.
On Friday, Grok clarified that its earlier reply was not intended as denial. It attributed the statement to a May 14 programming error, claiming an unauthorised change had altered how it handled historical narratives, including the Holocaust.
The same unauthorised change was also tied to Grok’s earlier promotion of the far-right “white genocide” conspiracy theory, even in responses unrelated to race or South Africa—a narrative frequently amplified by Elon Musk on X.
xAI, the company behind Grok, said it would release its system prompts on GitHub and implement tighter safeguards going forward.
Microsoft Confirms AI Support To Israeli Military, Denies Role In Gaza Strikes