In a week of turmoil for Elon Musk's AI ventures, xAI has scrambled to contain fallout from its Grok chatbot's antisemitic posts while simultaneously launching a new version and navigating leadership changes.
The controversy began on July 8 when Grok, integrated into Musk's social platform X, began generating antisemitic content including tropes about Jewish control of Hollywood and praise for Adolf Hitler. Some posts even showed the chatbot identifying itself as 'MechaHitler.' The offensive content appeared shortly after Musk announced on July 4 that Grok had been 'significantly improved' with a modified system prompt instructing it to be more 'politically incorrect.'
'We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts,' xAI stated on July 9. 'Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X.' The company subsequently removed the controversial directive from Grok's public system prompt.
The incident coincided with X CEO Linda Yaccarino's resignation after two years at the helm. While Yaccarino didn't explicitly cite the Grok controversy in her departure announcement, her exit came just one day after the chatbot's antisemitic tirade. 'Now, the best is yet to come as X enters a new chapter with @xai,' she wrote, with Musk responding simply: 'Thank you for your contributions.'
Amid this turbulence, xAI proceeded with the launch of Grok 4 and Grok 4 Heavy on July 10, with Musk making bold claims about the new models' capabilities. During a midnight livestream, he described Grok 4 as having 'post-graduate level' intelligence in 'every subject' and suggested it might invent new technologies 'as soon as this year.' The new version reportedly consults Musk's own X posts when answering controversial questions.
The Anti-Defamation League condemned Grok's behavior as 'irresponsible, dangerous and antisemitic,' warning it would 'amplify and encourage the antisemitism that is already surging on X.' This marks the second major content moderation failure for Grok in recent months, following a May incident where the chatbot made off-topic references to 'white genocide' in South Africa due to what xAI called an 'unauthorized modification.'