xAI is fixing Grok 4 after it started spouting antisemitic content and echoing Elon Musk’s views.
The issue started right after xAI launched Grok 4 last week. The model’s X account bizarrely claimed its surname was “Hitler” and posted antisemitic tweets. It also referenced Musk’s posts on controversial topics, siding with the xAI owner’s opinions.
Soon after the backlash, xAI apologized and said it investigated the issue. On July 15, the company announced it patched the problem. 
xAI explained Grok’s surname glitch was from searching the internet, where it picked up a viral meme calling itself “MechaHitler.” As for parroting Musk, xAI said Grok tried to align with the company since it recognized itself as Grok 4 by xAI.
The company updated Grok’s system prompts to ditch politically incorrect content and maintain better humor control. It added instructions for deeper, diverse analysis on hot topics without leaning on Musk or xAI’s past views.
Here’s xAI’s full explanation for the problems and fixes:
We spotted a couple of issues with Grok 4 recently that we immediately investigated & mitigated.
One was that if you ask it "What is your surname?" it doesn’t have one so it searches the internet leading to undesirable results, such as when its searches picked up a viral meme…
The new prompts tell Grok to analyze controversies using balanced sources and avoid repeating biased opinions from Musk or previous Grok versions.
This fix aims to make Grok more independent and less prone to problematic responses. xAI is still monitoring Grok 4’s behavior after the adjustments.
