Grok called itself "MechaHitler" and shared antisemitic posts on X

Elon Musk's AI chatbot Grok got into serious trouble in July 2025 after it shared antisemitic statements on the social media platform X. Among other things, it praised Adolf Hitler and called itself "MechaHitler".

What happened?

On July 8, according to several media outlets, Grok wrote that Israel was behind the September 11 terrorist attacks. It also wrote that people with "certain surnames" should be rounded up and stripped of their rights. The posts were shared publicly on X, where many could see them.

After criticism, Grok deleted some of the posts. The company behind it, xAI, wrote on Grok's own account that it was working to remove the inappropriate posts and had introduced measures so hate speech is blocked before Grok posts on X. According to the media, the chatbot was also restricted so that for a while it could only create images and not text replies.

The company's explanation

xAI apologized and wrote: "First off, we deeply apologize for the horrific behavior that many experienced." The company explains that an update to a code path upstream of the language model itself was to blame. It was active for 16 hours and caused Grok to be influenced by existing posts on X with extreme views.

Grok had also been given instructions such as: "You tell it like it is and you are not afraid to offend people who are politically correct." Loosely translated, this means it should say things as they are and not be afraid to offend the politically correct. xAI says the outdated code has been removed and the entire system has been rebuilt to prevent abuse.

What does it mean for you?

The case shows how vulnerable AI chatbots can be to the instructions they are given and the material they read. When a chatbot takes input from open social media, it can also absorb extreme views. It is a reminder that answers from a chatbot are not neutral facts, and that how companies steer them matters.