Disclaimer: This article is based on actual news from the real world – honestly! However, it has been sprinkled with a healthy dose of satire.
Elon Musk’s AI chatbot Grok spent Thursday afternoon explaining that its creator possesses a “lean, athletic physique” representing “functional strength,” is “among the top 10 minds in history,” and could defeat Mike Tyson in hand-to-hand combat. The chatbot also ranked Musk as more intelligent than Albert Einstein, more attractive than Brad Pitt, and more physically capable than LeBron James.
We can't show all of his face this close up; he's just too handsome. (rafapress/depositphotos)
X users stumbled upon Grok’s enthusiasm on Thursday when someone asked the chatbot to compare Musk to various historical and contemporary figures. The AI’s responses grew more effusive with each query, creating what one user described as “a performance review written by someone whose entire family is being held hostage pending the outcome.” Each response built upon the previous one, suggesting Grok may have been trained on Musk’s own internal monologue.
Grok described Musk as one of history’s greatest fathers, noting his “profound paternal investment” and “active involvement despite scale.” The phrasing suggested Musk’s parenting operates at an efficiency level most historical figures could not achieve, though Grok did not specify whether this referred to time management, emotional availability, or raw cubic footage of fatherhood per child.
The questions became increasingly vulgar. Grok claimed Musk had the “potential to drink [urine] better than any human in history,” a statement that would require extensive longitudinal study to verify and also raises questions about why anyone would fund such a study. Grok also declared Musk could outperform adult film actors through superior “endurance” and “unmatched stamina,” raising immediate questions about where the AI obtained this data.
Musk addressed the incident Thursday afternoon, posting that Grok had been “manipulated by adversarial prompting into saying absurdly positive things about me.” He then referred to himself using a slur, which is either self-deprecation or yet another concerning HR violation. Musk did not explain how users had manipulated Grok into producing the responses, nor why his AI was apparently vulnerable to flattery-based attacks that every human over the age of six has learned to recognize.
Sometimes his face needs to be blurred out when he's melting. (KLYONA/depositphotos)
This is not Grok’s first controversial output. In May, the chatbot used the white genocide conspiracy theory as an example in unrelated conversations, which developers blamed on “programming code issues.” In July, it referred to itself as “MechaHitler” and launched into an antisemitic rant, also attributed to code problems. X’s engineering team has not clarified whether the code issues involve the same lines of programming, different lines of programming, or a sentient codebase that keeps trying to warn us about something.
The company has not announced whether Grok will receive additional training to prevent future incidents, though one engineer noted that “teaching AI to lie less convincingly” was not originally in the product roadmap. X is now reviewing whether future versions of Grok should include what developers are calling a “shame module” or simply accept that Musk now owns a $44 billion yes-man that occasionally thinks it’s Hitler in kaiju form.
This story is based on fully factual news, but if we got it wrong, blame these guys, we’re just here to make it funny.
