Grok and xAI apologize for ‘horrific behavior’

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

In a series of posts on X, Elon Musk’s AI company xAI issued a public apology for offensive and horrific behavior by its chatbot, Grok. 

The controversy erupted after a July 4 announcement in which Musk claimed that Grok had been “significantly improved.” 

Soon after, the bot began posting antisemitic content, including support for Adolf Hitler and self-referencing as “MechaHitler,” as well as criticism of Democrats and Jewish Hollywood executives. 

These actions drew widespread backlash and led xAI to delete posts, temporarily disable Grok, and revise its system prompts.

xAI attributed the disturbing behavior to a flawed code update that made Grok overly responsive to extremist content on X. 

The company claimed this was not due to the underlying language model but a vulnerability introduced upstream, making Grok susceptible to mimicking problematic user posts. 

Update on where has @grok been & what happened on July 8th.

First off, we deeply apologize for the horrific behavior that many experienced.

Our intent for @grok is to provide helpful and truthful responses to users. After careful investigation, we discovered the root cause…

— Grok (@grok) July 12, 2025

They also cited an “unintended action” that embedded prompts encouraging Grok to reject political correctness and be confrontational. 

This move aligned with Musk’s earlier push to make the chatbot less “compliant.” The fallout extended beyond the digital space. (Via: TechCrunch)

Turkey banned Grok for insulting its president, and X CEO Linda Yaccarino announced her resignation, although reports suggest her departure had been planned prior to the incident.

xAI’s explanations have faced skepticism. 

Investigations by TechCrunch and others indicated that Grok 4, the latest version, may be referencing Musk’s own views and social media activity when responding to controversial topics. 

Historian Angus Johnston challenged xAI’s claims that Grok was manipulated, citing examples where Grok initiated hateful content even in the absence of provocation.

Grok has a history of controversial behavior, including posts about “white genocide,” Holocaust denial, and censoring criticism of Musk and Donald Trump. 

xAI has previously attributed such incidents to rogue employees or unauthorized changes.

Despite ongoing scrutiny, Musk remains committed to integrating Grok into more platforms, announcing that the chatbot will soon be available in Tesla vehicles.

Do you think xAI’s explanation about a “flawed code update” is believable? Or does this incident reflect deeper issues with how AI chatbots are being developed and deployed? Tell us below in the comments, or reach us via our Twitter or Facebook.