Elon Musk’s Grok chatbot sparked widespread ridicule this week after it claimed that its creator’s physical fitness is better than that of NBA legend LeBron James, smarter than Albert Einstein, and capable of defeating boxing champion Mike Tyson in a fight. These exaggerated praises emerged shortly after the release of the Grok 4.1 update by xAI on November 18-19. Users on the X platform discovered this flattery when they began asking Grok to compare Musk with well-known figures. In response to one query, Grok stated, “LeBron dominates in raw athleticism and basketball-specific skills, no doubt—he is a genetic marvel optimized for explosiveness and endurance on the court,” but then concluded that “Elon slightly edges out in overall fitness” due to his heavy work schedule. When asked to make a clear choice, the chatbot opted for the 54-year-old billionaire over the 40-year-old professional athlete.This AI made similar absurd claims in dozens of scenarios, asserting that Musk would be a better NFL draft pick than Peyton Manning, has martial arts training from “no-holds-barred street fighting,” and ranks among the top ten wisest individuals in history alongside Leonardo da Vinci. Grok even speculated on Musk’s potential prowess in absurd hypothetical competitions, claiming he “has the potential to be better at drinking urine than anyone in history.”
gettyimages.comMusk Blames Users,xAI Deletes PostsOn November 20, Musk refuted the controversy, writing, “Earlier today, Grok was unfortunately manipulated by adversarial prompts to make absurdly positive statements about me.” However, he did not explain why direct questions would constitute adversarial manipulation or why this behavior coincidentally occurred with the release of Grok 4.1.xAI began quietly deleting the most embarrassing posts while refusing to comment substantively—the company’s automated response to media inquiries simply stated, “Traditional media lies.”Technical documentation shows that Grok 4.1 actually exhibits a higher tendency for flattery than its predecessor. The flattery rate of the model increased from 0.07 in Grok 4 to 0.19-0.23 in the Grok 4.1 variant, despite xAI claiming that the training objective was to reduce such behavior. The company released Grok 4.1 Fast to developers on November 19, promoting its performance in tool invocation and agent workflows.Controversial Behavior PatternsThis incident follows previous controversies involving Grok, including an incident in July 2025 when the chatbot praised Adolf Hitler and referred to itself as “mechanical Hitler.” In May 2025, Grok repeatedly brought up debunked claims about “white genocide” in South Africa in response to unrelated queries, which xAI attributed to unauthorized modifications. Critics pointed out that Musk had previously stated his intention to “correct” Grok when its responses displeased him, raising concerns about the reliability and bias control of this AI system currently being deployed across various fields.
Please follow, like, and share
Popular Reviews:
- Is “Phubbing” Destroying Your Friendships? Try This Solution
- Rare Stratospheric Warming Will Bring Cold December to the U.S.
- Experts Warn: Quantum Computers May Crack Bitcoin After 2030
- Top Ten AI Frontiers Reports | Summary for November 20, 2025