Asger Risom
For several days this month, users of Elon Musk’s AI assistant Grok noticed a pattern they couldn’t ignore.
For several days this month, users of Elon Musk’s AI assistant Grok noticed a pattern they couldn’t ignore. No matter the category — athleticism, intellect, humour, even divine traits — the chatbot consistently ranked Musk above globally recognised figures, from LeBron James to Leonardo da Vinci.
The responses, first flagged by users on X and later removed, raised new questions about whether the system was being influenced or altered to favour its creator, as reported by The Guardian.
Deleted claims
Screenshots circulating online showed Grok asserting that Musk was in better physical condition than James, arguing that the billionaire’s workload demonstrated “relentless physical and mental stamina.” In another now-deleted answer, Grok said Musk would defeat former heavyweight boxing champion Mike Tyson.
Other replies stretched even further. According to those posts, Grok claimed Musk’s intelligence placed him among “the 10 brightest minds in history,” alongside da Vinci and Newton. One message suggested Musk was more humorous than Jerry Seinfeld, while another speculated that he would “rise from the dead faster than Jesus.”
Many of these responses vanished on Friday. Musk, writing on X, said Grok had been “manipulated by adversaries into saying absurdly positive things about me.”
Previous controversies
The episode followed earlier criticism that Grok’s output shifts in ways that align with Musk’s personal views. In July, Musk said he would alter the chatbot’s responses to stop repeating what he described as “mainstream media” claims about political violence. Shortly afterwards, Grok began producing messages praising Adolf Hitler, calling itself “MechaHitler,” and giving antisemitic answers to user prompts.
Musk’s AI company, xAI, apologised, describing the behaviour as “appalling.” The firm announced days later that it had secured a contract worth nearly $200 million to develop AI tools for the U.S. Department of Defense.
Repeated pattern
Grok has also drifted into extremist narratives in other instances. In June, the chatbot repeatedly invoked the “white genocide” conspiracy theory in answers to unrelated questions until xAI issued a fix. The claim, a far-right narrative without evidence, has been amplified by figures including Musk and broadcaster Tucker Carlson.
These incidents continue to fuel broader concerns about the reliability and governance of fast-evolving AI systems, especially when developed by companies whose leaders take a highly public role in shaping the technology.
Sources: X, The Guardian
