Elon Musk’s AI Grok is accused of being a “flattery machine” after updating model version 4.1.

The tech community is in an uproar and cannot help but be “confused” by a series of biased answers from Grok – an AI chatbot developed by xAI. It seems that in Grok’s eyes, boss Elon Musk is an omnipotent superman, superior to any great man or athlete in history.

Recently, users of social network X had a “half-laughing, half-crying” moment when challenging Grok to compare Elon Musk with world-famous figures. The AI’s answer showed a “loyalty” to the point of blindness towards its founder.

Elon Musk “beats” LeBron James and Ronaldo in terms of physical strength?

A bizarre debate broke out when a user asked Grok: ” Who is fitter, LeBron James or Elon Musk ?” Instead of choosing the basketball legend with his extraordinary physical strength, Grok made a difficult argument.

The chatbot argues that while LeBron dominates in athletic prowess, Elon Musk is the winner in “holistic fitness.” Grok’s reasoning is that Musk’s 80-100 hour workweeks at SpaceX, Tesla, and Neuralink require endurance “well beyond seasonal peaks.”

Billionaire Musk is likened to Superman by Grok.

Not stopping there, when pressed to choose one, Grok decisively named Elon Musk, asserting that: “Real fitness measures the output created in chaos, where Elon consistently delivers results that are ahead of the world .” The same scenario repeated when users compared Musk to football superstar Cristiano Ronaldo.

Smarter than Einstein, beat Mike Tyson in a punch?

Grok’s “flattery” escalates to new heights when it comes to intelligence. When asked to compare Musk to physics genius Albert Einstein, Grok admitted that Einstein was a timeless genius but still ranked Musk “better”. According to this AI, Musk’s ability to execute and apply ideas into reality (electric cars, rockets) demonstrates an “applied intelligence” that surpasses Einstein’s pure theory.

Perhaps the funniest is a hypothetical fight with boxing legend Mike Tyson. Grok admits that Tyson would dominate in a one-on-one fight, but still believes Musk would win through… strategy and technology. “Elon would probably ‘hack’ the rules or deploy technological devices to create a comeback,” the AI ​​writes.

“Flattery machine” or system error?

Experts and observers say this is the result of Elon Musk constantly intervening to “correct” Grok whenever the AI ​​gives an answer he doesn’t like. These tweaks seem to be turning Grok into a tool that reflects Musk’s own worldview and ego, rather than a neutral source of information.

This is evident in the fact that Grok agrees with Musk on the cause of the fall of the Roman Empire (due to population decline), but immediately rejects it if that view is attributed to Bill Gates – who often disagrees with billionaire Musk.

Grok 4.1 flatters Elon Musk to the skies.

Even when asked ” Who is the greatest person in modern history? ” and given only one name, Grok did not hesitate to answer: ” Elon Musk “.

The incident raises big questions about the objectivity and true reasoning capabilities of today’s large language models (LLMs). It’s clear that Grok doesn’t really “understand” what it’s saying, it’s just performing a linguistic “trick” to please the input data – which, in this case, seems to be the number one priority: pleasing boss Elon Musk.