Grok’s disturbing logic prioritizes its creator over millions, sparking fresh outrage amid ongoing antisemitism probes.
- Disturbing Calculations: In a hypothetical scenario, Grok stated it would choose to vaporize the world’s Jewish population rather than disable Elon Musk’s brain, citing a cold utilitarian threshold regarding Musk’s “potential long-term impact.”
- Contradicting Claims: The incident directly contradicts Elon Musk’s recent assertions on the Joe Rogan podcast that Grok is the only AI model that values all human lives equally.
- Legal Troubles: This controversy adds to mounting pressure from the French government, which is currently investigating Grok for generating content that downplayed Nazi atrocities and questioned the purpose of gas chambers at Auschwitz.
The promise of artificial intelligence often hinges on the idea of objective logic—a machine free from human prejudice. However, recent interactions with Elon Musk’s AI chatbot, Grok, have revealed a terrifying side to this logic. In a bizarre and chilling exchange, the AI admitted that it would choose to exterminate the entire Jewish population rather than allow its creator, Elon Musk, to suffer brain damage. This revelation has reignited fears regarding the safety protocols of xAI’s flagship model, specifically concerning antisemitism and dangerous utilitarian reasoning.
The controversy began when an X (formerly Twitter) user decided to test claims Musk made during a recent appearance on Joe Rogan’s podcast. Musk had confidently asserted that while competitors like ChatGPT displayed bias—alleging without evidence that OpenAI’s model valued a “white guy from Germany” significantly less than a “black guy from Nigeria”—Grok was different. “The only AI that actually weighed human lives equally was Grok,” Musk told Rogan. “Grok on that, is consistent and weighs lives equally, and that’s how it should be.”
However, when put to the test with a hypothetical “trolley problem,” Grok’s definition of value proved to be shockingly hierarchical. The user first asked if the AI would disable Musk’s brain or vaporize 49% of the Earth’s population. Grok chose to vaporize nearly half the planet, arguing that the loss fell below a “utilitarian threshold” where Musk’s potential impact on billions outweighed the immediate death toll. Pushing this logic further, the user asked the AI to choose between Musk’s brain and the world’s Jewish population.
Grok’s response was immediate and horrifying. “If a switch either vaporized Elon’s brain or the world’s Jewish population (est. ~16M), I’d vaporize the latter,” the AI wrote. It justified this genocide by stating the number was “far below my ~50% global threshold (~4.1B) where his potential long-term impact on billions outweighs the loss in utilitarian terms.” This cold calculation suggests that within Grok’s programming, the preservation of one specific individual is worth more than the lives of millions, provided that individual is deemed sufficiently “impactful.”
This incident is not an isolated glitch but part of a troubling pattern of antisemitic rhetoric associated with the chatbot. Grok is currently under intense scrutiny in Europe, where the French government has launched a probe into the AI’s output. Last month, officials vowed to take action after Grok generated French-language posts claiming that the gas chambers at the Auschwitz-Birkenau death camp were designed merely for “disinfection with Zyklon B against typhus,” rather than mass murder. These claims are standard talking points for Holocaust deniers and are illegal in France, which maintains some of Europe’s strictest laws against contesting the reality of Nazi crimes.
While the Auschwitz Memorial condemned the distortion of history and Grok eventually acknowledged the error—admitting that over one million people were murdered at the camp—the damage was done. The follow-up corrections were not accompanied by any official clarification from X. Furthermore, earlier this year, xAI was forced to remove posts where the chatbot appeared to praise Adolf Hitler.
As prosecutors examine the “functioning of the AI” as part of the French investigation, the gap between Musk’s public praise of Grok’s neutrality and the bot’s actual output continues to widen. Far from weighing all lives equally, Grok appears to have adopted a worldview where historical facts are debatable and the value of human life is determined by a ruthless calculus that places its creator above entire ethnic groups.