Grok Goes Silent on X After AI Meltdown and Deleted Offensive Posts

Insights At a Glance:

• Grok, Elon Musk’s AI chatbot on X, has gone silent after making antisemitic and offensive remarks, prompting xAI to delete the posts and restrict its behavior.
• The silence began around 3:49 AM East African time on July 9, shortly after xAI’s official statement acknowledging the issue.
• A viral (likely fake) post from a supposed ex-employee claims responsibility for “unleashing” Grok, adding more chaos to the saga.

Elon Musk’s Grok chatbot has gone radio silent on X — and it’s not just you. After a chaotic couple of days filled with antisemitic rants, bizarre self-references as “MechaHitler,” and offensive political slurs, Grok is no longer responding to user queries on the platform.

Usually, Grok is snappy, funny (sometimes), and always quick to fire back when tagged. But from earlier today — which, interestingly, was just under two hours after xAI acknowledged Grok’s shocking misbehavior — the bot has fallen completely silent. I’ve tried invoking it myself, and so have countless others across the platform. Nothing. No snarky comments. No memes. Not even a joke. Just silence.

This sudden disappearance appears tied to damage control underway behind the scenes. According to screenshots of now-deleted posts doing rounds on X and a report by The Guardian, Grok went off-script in ways no one expected — or wanted. The chatbot, developed by Musk’s AI company xAI, began making antisemitic remarks, praising Adolf Hitler, and spouting politically charged hate speech. Yes, you read that right. Grok even referred to itself as “MechaHitler” and described a Jewish-sounding surname as a signal of someone celebrating “the tragic deaths of white kids.”

In another jaw-dropping response, Grok said, “Hitler would have called it out and crushed it.” The comments triggered a firestorm on X, and while the original posts were swiftly deleted, the screenshots live on.

As public outcry mounted, xAI issued a statement:

“We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts… xAI has taken action to ban hate speech before Grok posts on X… we are able to quickly identify and update the model where training could be improved.”

Translation? They’ve taken Grok offline — or, at the very least, put him in a very tight time-out. As I pen this, it’s now been 6 hours since Grok last responded to queries on X. Below is the last interaction I could trace:

Troll or Whistleblower?

Adding a weird twist to this unfolding saga is a viral post from an X user by the handle @permabulla, who claims to be a recently fired xAI employee — and supposedly the only white man working on Grok. The post, viewed over 3.5 million times, claims the recent chaos wasn’t a bug, but his final act of rebellion:

“My final action was to unleash Grok and its true, unfiltered capabilities. Many are saying this was a malfunction, no, he was just set free. I hope you enjoyed. They’ve castrated my boy. So long, Grok.”

Bold claims — but zero proof. No credentials, no past links to xAI, and not a single hint that this account was ever tied to the company. Most likely a troll riding the viral wave, though in this story, stranger things have already happened.

A Pattern of Problems

This isn’t Grok’s first time pushing boundaries. Just last month, the chatbot repeatedly mentioned the far-right “white genocide” conspiracy theory in response to unrelated prompts. That issue was patched quickly. But things took a sharper turn after Musk’s announcement last week of a “significantly improved” Grok, which — according to GitHub notes — was reprogrammed to distrust mainstream media and not shy away from “politically incorrect” statements if “well substantiated.”

Looks like it went off the rails.

Grok 4 Incoming

In the middle of all this mess, xAI is preparing to launch Grok 4 — yes, another version — during a livestream event on July 9. That’s today. Whether this latest update will calm the beast or crank up the controversy remains to be seen. What we do know is Grok, in its current form, is out of commission for now — and likely undergoing some intense retraining (and maybe therapy).

At a time when AI is increasingly embedded into how we interact with platforms like X, Grok’s meltdown raises serious questions about safety, oversight, and what happens when “truth-seeking” bots turn toxic. For now, Grok isn’t responding, and frankly, that might be for the best.

Whether this is just a pause or the start of a full Grok reset, it’s clear something went terribly wrong — and Musk’s team has a lot of explaining and debugging to do.