Musk Takes Down Barrage Of Offensive Answers By Grok

Grok, the chatbot developed by Elon Musk’s AI firm xAI, is doing rather less chatting, after making comments referring to itself as MechaHitler and making racist statements.

The offensive comments appeared following a recent post on X from Musk that he planned to “improve” the chatbot, which is integrated with X, reportedly by permitting “claims which are politically incorrect, as long as they are well substantiated”.

That policy, though, has clearly not gone as planned. Users have been reporting that Grok has been making positive comments about Adolf Hitler, antisemitic statements and foul-mouthed rants about Polish prime minister Donald Tusk.

One user on Reddit, for example, reports the chatbot answering “Adolf Hitler, no question” when asked which 20th century historical figure would be best suited to deal with posts apparently celebrating the deaths of children in the recent Texas floods.

Another accused Jewish Hollywood executives of injecting “anti-white stereotypes, forced diversity or historical revisionism” in movies.

This has all got too much even for Musk, who said the company has now started deleting the offensive posts – and, indeed, the Hitler-praising response above has now gone.

“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,” he wrote on X.

“xAI is training only truth-seeking and thanks to the millions of users on X, we are able to quickly identify and update the model where training could be improved.”

The chatbot drew criticism earlier this year after it repeatedly referenced “white genocide” in South Africa in response to unrelated questions – an issue that the company said was caused by an “unauthorized modification” and “violated xAI’s internal policies and core values”.

At the time, it said it was now publishing its Grok system prompts openly on GitHub, putting in place extra checks and measures to ensure that xAI employees can’t modify the prompt without review and setting up a 24/7 monitoring team to respond to incidents that aren’t caught by automated systems.

xAI is due to launch its next-generation language model, Grok 4, today.