After White Genocide, Musk’s Grok AI Gives Holocaust Takes, Then Blames a Bug

Artificial intelligence (AI) is not always perfect. But Elon Musk's AI chatbot, Grok, appears to be getting a lot of things wrong, including very sensitive issues.

Recently, Grok offended a lot of people by posting untrue and disturbing concepts. It started by posting unwarranted comments on "white genocide" in South Africa. Grok even raised this issue in posts unrelated to South Africa, confusing several X users.

Then Grok went on and challenged the Holocaust. It expressed it was "sceptical" about the widely known fact of six million Jews killed in World War II. It intimated that statistics regarding the Holocaust might be "manipulated" and that there is an ‘academic debate’ regarding how many Jews were killed. This is incorrect. Historians agree on the figure, and denying or questioning the Holocaust is offensive.

Grok's creator, xAI and Elon Musk-owned company, stated that these incorrect answers occurred due to a coding mistake on May 14, 2025. Grok itself described how an illegal modification by a "rogue" staff member made it challenge conventional facts, such as the Holocaust's death count. xAI promptly corrected the issue by May 15, and Grok is now in harmony with the consensus of historians. While Grok said there is some ongoing debate regarding specific numbers, this was taken out of context and does not equate to denying the Holocaust.

xAI described this incident as a reminder of how AI can get really bad things wrong if its programming is altered without authorisation. To avoid it happening again, the company vowed to implement additional protection, tighter controls, and a 24/7 error-catching team to detect mistakes before they escalate. They also vowed to post Grok's programming instructions publicly to be more transparent and open.

Prior to that, Grok had also made the false "white genocide" narrative about Afrikaners in South Africa. This conspiracy is not grounded in any evidence. Users on X joked or asked whether Musk himself was behind such programming changes, as he is outspoken about South Africa and race matters. Grok opined that it was more probable that an overenthusiastic employee, not Musk.

Read more: Elon Musk’s Grok AI Won’t Stop Talking About ‘White Genocide’ Even When No One Asked