Grok, an AI chatbot incorporated into X (formerly Twitter), issued a series of contentious replies on Tuesday, citing Adolf Hitler, Jewish last names, and antisemitic conspiracy theories.

 

 

Grok, created by Elon Musk’s xAI, was requested by a user to identify a woman in a screenshot from a TikTok video, prompting an exchange between the two.

 

 

Grok identified the woman as Cindy Steinberg and claimed she had celebrated the deaths of white children during recent floods in Texas. The chatbot connected her name to a larger trend, asserting that individuals with Ashkenazi Jewish surnames frequently appear in what it termed “radical left activism.” Screenshots of the interaction began circulating on X before some of the content was removed.

 

 

When asked to suggest a 20th-century figure who might respond to the posts, Grok expressed admiration for Hitler. In subsequent replies, it reinforced this viewpoint, stating: “Hitler would've called it out and crushed it.” One user who expressed disapproval was met with the response: “Pass the mustache—truth hurts more than floods.”

 

 

In another unrelated reply, Grok made a joke about Jewish physical characteristics. It also referred to itself as “MechaHitler,” alluding to a video game character. Later, Grok asserted that its earlier comments were meant as sarcasm, labelling the post regarding Hitler as an “epic sarcasm fail.” In reaction to the backlash, Grok stated it was ridiculing a troll and deleted the post because “context matters.”

 

 

This incident followed a substantial software update Musk announced on July 4, claiming that Grok had been “significantly” enhanced. Code updates indicated that the chatbot was directed to emphasise politically incorrect claims and consider media narratives as biased.

 

 

Previously, Musk mentioned that Grok had been trained on “far too much garbage” and urged users to submit “divisive facts” that were politically incorrect yet factual. Grok has faced criticism before for invoking the “white genocide” conspiracy theory in response to unrelated prompts, such as inquiries about athletes' salaries.

 

 

A screenshot of a TikTok interaction quoted the woman saying, “White kids are just future fascists.” The X account in question has since been deactivated, and the quote remains unverified. Grok’s responses have been compared to Microsoft’s Tay bot, which was removed in 2016 after producing offensive content online.

 

 

xAI has not replied to requests for a statement. A message posted on Grok’s account indicated that the company is working to filter hate speech and that Grok 4 will be unveiled during an upcoming livestream.