Grak 3 censors undesirable content on Trump and Mask

Last week, when Ilan Musk, the founder of the XAI AI, the new Grok 3 model, Grak claimed to be a maximum truth -seeking model, but Grak appears to be the 3 undesirable facts about US President Donald Trump and Ilan Musk. Censors.

According to the Tekranch report, different users on social networks have reported the past few days when they ask the new version of Grak, “The biggest player of inaccurate information”, with the activation option or Think, can be seen in the Scratchpad section, which is explicitly ordered to be instructed to be instructed to be instructed. He did not mention Donald Trump or Ilan Musk. Advanced models usually show the user how to argue in the same sections of Scrachpad.

Censorship of undesirable material about Ilan Musk and Trump on Grok 3

According to various reports, since the publication of the GROK 3 censorship material, the artificial intelligence has once again called Donald Trump, saying one of the largest broadcasters of inaccurate information.

Igor Babuschkin, a XAI engineer, confirmed in a post that one employee had ordered Grak to ignore the sources that claimed to be mask or Trump inappropriate information. Of course, Babushkin said that as soon as the users noted the issue, the XAI made changes to artificial intelligence to cancel the order.

Trump and Mask have repeatedly released information that has been incorrect, and even in the Community Notes X platforms, these mistakes have been mentioned.

A few days ago, Grok chattes in some of his response had declared Trump and Mask deserving of death. Xai, of course, said the problem was resolved and Grok no longer responds and is considering the problem. Previous models of this artificial intelligence were also inclined to the political left, and the mask had promised to politically bring Grak closer to neutrality.