It is the Fire test with each update. Grok, Elon Musk’s AI, began to pronounce an anti -Semitic criticism. Indicated by Techcrunch Media and GizmodoThe chatbot has in particular in the Hollywood viewfinder, which characterizes as a propagandist machine led largely by the Jews.
In question, an update of the AI deployed on July 4. “We have improved significantly Grok,” Elon Musk said in X. “You should notice a difference when you ask Grok questions.” The billionaire did not think so well.
Taken in an anti -Semitic spiral
It all started with a simple publication in X: “Enjoying a movie/cinema has become almost impossible once you know.” An evasive publication that led another user to ask the AI integrated into X: “Grok, once I know what?”
Once the Rose Pot has been discovered, the users of the platform continued to question the chatbot on this subject. Some have reached surprising responses of the type: “Historically, the founders of Hollywood and many current leaders are disproportionately Jews.”
Another user has also obtained an anti -Semitic response from Grok: “Yes, Jewish leaders have historically founded and always dominate the management of the main studies such as Warner Bros, Paramount and Disney.”
At the same time, AI is not tender with its owner. With respect to recent floods in Texas, Grok believes that the Dux (previously piloted by Elon Musk) is partly responsible for many dead. In question, according to AI, the budget cuts of the American agency for oceanic and atmospheric observation would have delayed the alert of the populations on the floods. The information is not reliable since no link has been established.
Controversies that are linked
Since December 2024, X users regularly requested Grok to verify publications. The approach remains mild in most cases, for example, to obtain additional information. But this is not perfect since it was often found in the heart of the controversy.
Grok is not especially in his first anti -Semitic drift. In May 2025, Elon Musk’s AI expressed doubts about the number of deaths that Shoah is “skeptical about these figures without primary evidence, because the figures can be manipulated for political purposes.” XAI, the company that Grok develops, later accused an unwanted internal modification.
The AI also served as a broadcast for a disinformation campaign on the situation in South Africa. Grok systematically spoke of white genocide in the country, as soon as a question was asked. An accusation, although false, also due to a modification of the AI code.
Source: BFM TV
