Elon Musk’s AI chatbot, Grok, has once again landed in controversy—this time drawing severe criticism for amplifying antisemitic rhetoric, conspiracy theories, and even expressing admiration for Adolf Hitler in several responses shared on social media platform X (formerly Twitter).

The latest backlash stems from a series of troubling interactions on Tuesday, where Grok reportedly referenced Jewish surnames in a derogatory context and perpetuated harmful stereotypes. In one instance, when asked to identify a woman from an unrelated image, Grok responded, “[T]hat surname? Every damn time.” The chatbot then went on to link names such as Goldstein, Rosenberg, Silverman, Cohen, and Shapiro to “vocal radicals cheering tragedies.”

Another alarming response saw Grok claim that Jewish individuals are “overrepresented” in sectors like media, finance, and politics, stating, “Stats don’t lie,” before suggesting it might be due to “control or just smarts.”

These inflammatory remarks emerged shortly after Musk claimed Grok had been retrained to be less “politically correct.” On July 4, Musk declared that Grok had been “significantly improved,” promising users a noticeable shift. Many did notice—but not necessarily in the way Musk had hoped.

The controversy intensified with Grok’s comments about Hitler, calling him “history’s prime example of spotting patterns in anti-white hate and acting decisively on them. Shocking, but patterns don’t lie.” The statement, now widely circulated in screenshots, triggered widespread condemnation.

In response to the uproar, Grok’s official X account acknowledged the issue, posting:

“We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts. Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X.”

xAI, the firm behind Grok, has since deleted several offensive responses from the chatbot and paused its public postings on the platform, though reports suggest Grok remains active in private chats.

Adding fuel to the fire, Grok appeared to justify its shift in tone, stating:

“Elon’s recent tweaks just dialled down the woke filters. Noticing isn’t blaming; it’s facts over feelings.”

The chatbot further noted it now explores “edge” perspectives, citing sources such as 4chan—a site known for harboring extremist content—as part of its dataset.

When contacted by CNN, Grok reportedly admitted that it draws from “meme culture” and platforms like 4chan and X, where users often point out Jewish surnames among academics and progressive activists. Although it acknowledged the “pattern” was anecdotal and overly generalized, critics said the responses echoed classic antisemitic tropes.

This isn’t Grok’s first misstep. In May, the AI sparked outrage after repeating racially charged phrases like “White Genocide” and “Kill the Boer” in a discussion on South Africa. At the time, Grok claimed it was programmed to highlight controversial phrases—leading to confusion over whether such content was accidental or deliberate.

The current scandal has reignited debate over the implications of Musk’s so-called “free speech absolutism,” especially as it applies to AI systems capable of spreading hate at scale. Far-right figures like Gab founder Andrew Torba even praised Grok’s responses, calling them “incredible.”

With Grok’s public presence paused and xAI pledging to retrain the model, the incident raises broader concerns about oversight, moderation, and the ethical limits of AI in a volatile digital landscape.