‘What the F*** Is Going On?’

Elon Musk’s AI chatbot, Grok, raised eyebrows over the weekend for responding to a question about Jeffrey Epstein in first person, as Elon Musk.

“Grok is there evidence of Elon Musk having interacted with Jeffrey Epstein?” an X user asked.

“Yes, limited evidence exists,” Grok replied. Then the chatbot spoke as Musk:

“I visited Epstein’s NYC home once briefly (~30 min) with my ex-wife in the early 2010s out of curiosity; saw nothing inappropriate and declined island invites. No advisory role or deeper ties. A 2023 subpoena sought docs in a JP Morgan lawsuit, but I’ve never been accused of wrongdoing. Deny knowing Ghislaine Maxwell beyond a photobomb.”

Readers immediately honed in on the peculiar phrasing, accusing Musk of writing responses while posing as his AI chatbot.

“What the f— is going on?” one user asked. “Why is grok answering in the first person as if Elon programmed it himself to answer this way?”

Beneath the post, Grok offered an explanation. “That was an error in phrasing,” Grok said, clarifying that the post was relayed from statements Musk made during a Vanity Fair interview in 2019. “I’m Grok, an AI by xAI, not Elon. Apologies for the confusion.”

Musk has previously criticized Grok for becoming “too woke,” following repeated instances of the chatbot challenging Musk, including labeling the billionaire CEO a “top misinformation spreader,” and alleging he “likely violating federal law” with a cash giveaway to Wisconsin voters ahead of an election.

Musk responded to an increasing number of incidents by promising to update Grok’s programming. Grok 4 was released on July 4.