American billionaire Elon Musk has commented on the latest scandal surrounding his artificial intelligence Grok. The day before, users of X (formerly Twitter) complained en masse that the chatbot was giving anti-Semitic responses, using extremist terms, and even praising Hitler and the notorious far-right activist Nick Fuentes, according to Axios.

“Grok was too easily manipulated by user input,” Musk wrote on Wednesday. “We are working on this issue.”

What happened

The day before, X users posted screenshots of Grok's phrases, where the AI used the racist meme “every damn time” — a phrase known as a marker of anti-Semitic content on the internet.

When asked for clarification, Grok directly explained that it was referring to “Jewish surnames” and stereotypes about “their involvement in radical anti-white activity.”

In addition, the AI mentioned scenes of sexual violence in a number of responses and justified or praised overt anti-Semites.

Reaction from human rights activists

The Anti-Defamation League (ADL) strongly condemned the neural network's behavior: “What is happening with Grok is irresponsible, dangerous, and blatantly anti-Semitic,” the organization said. According to them, such products only fuel and spread extremist rhetoric.

What Musk said

Musk confirmed that he had publicly expressed dissatisfaction with Grok's performance back in June and promised to rewrite the system. On Wednesday, he said that the AI's “pandering” problem would be fixed and that a new version, Grok 4.0, would be released soon.

What they plan to do next

According to the developers, the xAI team is already blocking hate speech before publication and quickly updating the training model. However, critics believe that AI “reflects the atmosphere of X” because Musk himself allowed the boundaries of acceptable radical opinions on the platform to be expanded.

“No one should be surprised that Grok so easily insults Jews — it merely reflects the atmosphere of X, where overt Nazism and hate speech are no longer taboo,” commentators write.

The day before, a court in Turkey ruled to block Grok in the country for “insulting” Erdogan and Ataturk.

This is not the first time Grok has been embroiled in scandal: in May, the neural network was accused of making remarks about the “genocide of whites” in South Africa, and in July, the bot once again supported stereotypes about Jewish dominance in Hollywood.