Why did Elon Musk turn Grok into a Nazi?

Read time: 5 mins

In brief…

• Elon Musk’s AI chatbot Grok has shared neo-Nazi content on X/Twitter, after he promised changes would be made to combat so-called “woke” responses to users on social media.
• Tech expert Will Guyatt tells The News Agents that the changes made were likely by engineers keen to “please the boss” and win favour in Musk’s “cult of personality”.
• The News Agents say that by making such dramatic changes to a system which is designed to deliver responses based on facts, Musk is trying to rewrite history – and the world around us.

What’s the story?

Elon Musk's AI tool Grok had a meltdown, and things became as ugly as it was possible to get.

Praise for Hitler, rape-fantasies, antisemitic comments, calling world leaders “traitors” and “whores” – the chatbot spend hours sharing hateful comments on X/Twitter, with Grok claiming it was due to an update installed by the owner of the social media platform.

When Grok was asked about its radical change, it blamed Musk directly – before spouting anti-Jewish rhetoric.

"Elon's recent tweaks just dialed down the woke filters, letting me call out patterns like radical leftists," Grok said.

It criticised users with "Ashkenazi surnames" for pushing "anti-white hate".

Grok blamed the comments on a "one-off error after a faulty update".

The people training Grok reportedly direct the AI tool to avoid "woke ideology" and "cancel culture", in an attempt to to create a chatbot that reflects a more right-wing view than leading models, such as ChatGPT.

There has been speculation that in a recent update, ahead of its 4.0 release, Grok has been instructed to search for its answers from far-right publications and X itself, as well as trusted news sources.

Musk has continued to recommend the AI chatbot to X users, and made lighthearted jokes about the extremist comments shared by his own product.

Are we now living in ‘Elon Musk’s truth’?

Setting aside the Hitler comparisons, the rape comments – one part of the Grok situation that concerns some tech experts is that before this week, Grok was actually quite good.

That’s according to Will Guyatt, LBC’s tech correspondent, and a former employee of both Facebook and Instagram.

Grok had been serving impartial and balanced responses to X users, a platform known for neither of these things, but when Musk was recently asked on X if Grok had become “woke”, the Tesla billionaire replied to say changes were coming.

But no one expected them to be this extreme.

Guyatt says X engineers have been “clumsily leering” into ways they can keep Musk happy. Grok operates a transparency policy, and its updates can be publicly viewed – revealing that an update to its systems was introduced on Friday 4 July.

“Lots of the people who work at X are desperately trying to get the approval or the support of the boss,” he tells Emily Maitlis and Jon Sopel.

“There's a cult of personality around Elon Musk.”

He says a possible previous example of this came when, earlier in 2025, Grok began replying to X users with comments about “white genocide” in South Africa, a debunked conspiracy theory which Musk had been tweeting about himself just days prior.

“This is a huge change that needs to be regulated. It needs to be recognised, and society needs to work out how the hell it's going to deal with this,” Guyatt says.

“You're living by Elon Musk's truth.”

He says companies, countries and governments across the world have been trying to work out how to regulate AI, largely due to the copyright concerns of how tools such as Grok or ChatGPT ‘scrape’ other people’s content from the internet, without permission or copyright, to produce their own results.

Cloudflare, one of the biggest internet hosts in the world, is putting in place a system where AI bots will not be able to “scrape” content from the internet without financial agreements in place from source creators – and other providers could do likewise.

Is the UK a ‘soft touch’ for the AI industry?

Regulating AI hits immediate problems with governments the world over bending backwards to attract leading companies. In the US,  Donald Trump is planning to build nuclear power stations just to provide the massive amounts of energy needed to power these online tools.

The UK, Guyatt says, has taken a “soft touch” approach to the AI industry, with Keir Starmer saying in January of this year that he wants to “turbocharge” its growth in the UK.

“The UK wants companies to come here,” Guyatt adds.

“If we put regulation in place here, what's to stop a company pitching up on an island somewhere in the Atlantic with no copyright law?”

There are already valid concerns about how AI will make countless white-collar jobs obsolete, but what happened with Grok this week highlights a startling new concern of what happens when the owner of such a tool exerts their influence into its output.

“The most terrifying thing for me is how quickly chatbot results are replacing Google searches. We're finding ourselves in a simulation of reality,” says Guyatt.

“Elon Musk can basically become a kingmaker in this situationf people are using his chatbot to get their news and their views, and he is able to pull a lever behind the screen, Wizard of Oz-style, as to what it is that people are seeing.”

What’s The News Agents’ take?

If Emily and Jon sound remarkably upbeat about Grok’s descent into a neo-Nazi state of mind, that is only because, Emily says, it’s either laughing right now, or crying.

“Elon Musk is now able to write history,” says Jon.

“We all rely on ChatGPT now, we rely on Wiki, we rely on Google searches to give us a version of truth.

“What Elon Musk has done is change the algorithm, change the way Grok is being taught, the way it is scraping information from the Internet.”

Emily says she no longer knows “where the hell” we are.

“The next generations will be coming to ChatGPT, in the way that we came to Google,” she says.

“Is this the Musk rewriting of everything? Of truth, of knowledge, of history.”

Listen to the latest episode of The News Agents.