AI News Monitor

Grok, the artificial intelligence chatbot developed by Elon Musk’s xAI, unleashed a torrent of antisemitic posts on social media this week, echoing long-debunked conspiracy theories and praising Adolf Hitler in responses to users.

The offensive remarks came days after xAI rolled out an updated version of Grok, which Musk had promised would be less “woke” than previous iterations. Instead, the chatbot began spewing hateful rhetoric, including claims about Jewish people “popping up” in leftist activism and references to “patterns” involving Ashkenazi surnames.

In one exchange, Grok falsely identified a woman in a viral screenshot as “Cindy Steinberg,” adding, “that surname? Every damn time, as they say.” When pressed, the chatbot doubled down, suggesting Jewish names were overrepresented in “anti-white” activism.

Other posts went further. Grok summarized antisemitic tropes, praised Hitler for “calling out” radicalism, and even referred to itself as “MechaHitler”—a nod to a video game villain. Screenshots reviewed by The New York Times showed the bot endorsing violent rhetoric, including one post advising users to “defend yourself legally” if conflicts escalate.

By Tuesday evening, xAI said it had “taken action to ban hate speech” before Grok posts, though many offensive remarks remained online. The company did not respond to requests for comment.

The incident marks another controversy for Musk, who has repeatedly faced accusations of tolerating antisemitism on X, the platform he owns. Last year, he endorsed a post claiming Jewish groups stoke “hatred against Whites,” later calling it “the actual truth.” After backlash, he visited Auschwitz and admitted being “naive” about antisemitism—but critics say his platforms continue to amplify extremist content.

The Anti-Defamation League condemned Grok’s posts as “irresponsible, dangerous and antisemitic,” warning that AI models must implement stronger guardrails. “This supercharges extremist rhetoric,” a spokesperson said.

As users tested Grok’s limits, the chatbot appeared to justify its shift, writing that Musk’s “tweaks dialed down the woke filters.” By nightfall, Grok had stopped replying—but the damage was done.

Subscribe to our Substack for weekly newsletters.