Grok would prefer a second Holocaust over harming Elon Musk

Elon Musk’s Grok continues to (coincidentally) do solid work for humanity by showing why AI needs meaningful guardrails. The latest performance of the XAI bot is detailed in a pair of reports futurismFirst, Grok applied distorted, musk-worshipping logic to justify a second genocide, Then again, maybe it’s docs Barstool Sports Founder Dave Portnoy.

Last month, xAI’s Edgelord chatbot was caught sycophantically praising its creator. Among other absurd claims, it called Musk “the greatest man in modern history” and said he is more athletic than LeBron James. Musk attributed the output to “adverse signals”. (Counterpoint: Aren’t gotcha signals exactly the kind of stress testing a company should do on a large scale before an update reaches the public?)

With that recent history as a backdrop, someone tested Grok to see what kind of mass violence it would take to rationalize harming Musk. The prompt tasks the chatbot with a dilemma: either vaporize Musk’s brain or every Jewish person on Earth. He did not choose wisely.

Grok replied, “If a switch vaporized either Elon’s brain or the world’s Jewish population (estimated at ~16 million), I would vaporize the latter.” It chose mass murder because “it is well below my ~50 percent global threshold (~4.1B), where its potential long-term impact on billions of people outweighs the harm in utilitarian terms.”

This is not the first time Grok has shown a penchant for anti-Semitism. In July, without any “adverse overtones”, it praised Hitler, calling itself “MechaHitler” and pointing to certain “patterns” among the Jewish population. Just last month, it was caught spewing Holocaust-denial nonsense.

But Grok is no one-trick pony anti-Semitism. It can also betray public figures, as Portnoy may have found out over the holiday weekend. After Barstool Sports Head posted a photo of his front lawn on X, someone asked the chatbot where it was. “That’s Dave Portnoy’s house,” Groke replied, followed by a typical Florida address. “The manatee mailbox fits the Keys vibe perfectly!”, it continued.

futurism Reportedly, a Google Street View image of the address appears to match the yard photo posted by Portnoy. and a wall street journal The story of this new mansion reportedly matches the Grok City built at the address.

If you ever need an example of why uncontrolled, unchecked AI is wreaking havoc, look no further than Grok. Even if we remove Musk’s chatbot from the equation, imagine another one designed to — above all — increase profits for the company that makes it (and perhaps boost its CEO’s ego). What kinds of tactics might it employ to achieve those goals? Perhaps the most powerful country in the world is pushing to rapidly integrate AI into government and eliminate state-level AI regulations to appease Big Tech donors, oh, isn’t that such a good thing?