Grok goes Nazi, Elon Musk’s AI praises Hitler

Elon Musk, owner of xAI, had mentioned on the X platform a “significant improvement” in his chatbot, without specifying the changes in its operation.

Many of Grok’s shocking replies have been deleted since they were posted, but Internet users have nevertheless published screenshots of the conversational agent’s replies that illustrate its new editorial line.

In some responses, Grok evoked “anti-white stereotypes”. In others, it called historical Hollywood figures “disproportionately Jewish”, and Israel “that tacky ex who still complains about the Holocaust”. He then praised Hitler, even sometimes calling itself “MechaHitler”.

The American NGO Anti-Defamation League (ADL) described this type of content as “irresponsible, dangerous and anti-Semitic”.

The AI tool has also attacked heads of state, such as Turkish President Recep Tayyip Erdogan. Grok called him a “snake”. In response, an Ankara court on Wednesday blocked Turkish Internet users’ access to Grok and opened an official investigation into the incident.

The Polish authorities have also denounced Grok’s offensive remarks about their country’s political leaders, including Prime Minister Donald Tusk. They have even reported xAI to the European Commission.

In a tweet published on July 9, the teams responsible for Grok said they would remove the chatbot’s recent problematic posts as they came.

Elon Musk didn’t comment directly on the controversy, but did post a message on X that might allude to it: “There’s never a dull moment on this platform.”