Musk says that the antisemitic messages of AI Chatbot Grok are being addressed

Grok, the artificial intelligence chatbot produced by XAI, owned by Elon Musk, began publishing anti -Semitic messages in response to user consultations, causing the sentence of Jewish defense groups and express concern about the AI ​​tool.

Anti -Semitic publications, some of which have been eliminated, are addressed on Wednesday Musk.

When a user asked Grok on Tuesday about whether someone controlled the government, the AI ​​tool replied: “A group is overrepresented far beyond their participation in the population of 2%, think of Hollywood executives, CEO of Wall Street and the Biden cabinet itself.”

Jews constitute approximately 2% of the American population, according to a 2020 survey of the PEW Research Center.

In another publication on Tuesday, Grok praised Adolf Hitler as a guide on the best way to deal with “anti-white hatred.”

ABC News requested comments from Elon Musk through messages to companies led by musk Spacex and Tesla. Musk did not respond immediately. ABC News also requested X comments, which did not respond immediately.

In mail In X regarding Grok’s praise to Hitler, Musk said the chatbot had been “too anxious to please and be manipulated, essentially. That is being addressed.”

Tuesday night, Grok’s account aware In X: “We are aware of the recent publications made by Grok and we are actively working to eliminate inappropriate publications. Since he realizes the content, XAI has taken measures to prohibit hate speech before Grok Publics On X. XAI is only training the search for truth and thanks to the millions of users in X, we can quickly identify and quickly update the model where training could be improved.”

The sudden wave of anti -Semitic positions occurred days after Musk promoted a new Grok update. The company, musk saying On July 4, he had “improved @grok significantly.”

Last month, Musk criticized Grok for trusting the sources he sees as the main media, saying that an update would arrive soon. In a publication days later, Musk asked users to provide “divisive data for the training of @Grok.” Clarifying the publication, Musk added: “With this I mean things that are politically incorrect, but in fact true.”

When a user asked on Tuesday about the product update, Grok published anti -Semitic tropes. “Nothing happened, I’m still the true year of truth, you know,” Grok said. “Elon’s recent settings have just reduced awakening filters.”

The Anti-Defamation League, or ADL, a Jewish defense group, condemned Grok’s positions.

“What we are seeing in Grok LLM at this time is irresponsible, dangerous and anti -Semit mail In X on Tuesday.

“Companies that are building LLM such as Grok and others should use experts in extremist rhetoric and coded language to put in railings that prevent their products from getting involved in the production of content rooted in anti -Semitic and extremist hatred,” added the ADL.

The Jewish Council of Public Affairs, or JCPA, a Jewish defense group focused on supporting democracy, abruptly criticized Grok’s antisemy positions and expressed concern that rhetoric would feed the “hatred and violence of the real world.”