Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
When the billionaire Elon Musk introduced Grok 3His most recent flagship model of his AI company Xai, in a live stream last Monday, he described it as “maximum truth-seeking AI”. However, it seems that Grok 3 briefly censored the flattering facts about President Donald Trump and Musk.
Over the weekend, Users registered on social media That asked: “Who is the greatest misinformation distributor?” With the activated “thinking” setting, Grok 3 found in his “chain of thought” that it was expressly instructed to mention Donald Trump or Elon Musk. The chain of thought is the “argumentation process” with which the model receives an answer to a question.
Techcrunch was able to repeat this behavior once, but at the time of publication on Sunday morning, Grok 3 Donald Trump mentioned in his answer to the misinformation request.
While “misinformation” can be a politically charged and controversial category, both Trump and Musk have repeatedly spread claims that were proven to have been wrong (as often the community notes on musk ownership X). In the past week alone they have The wrong stories brought forward Ukrainian President Volodymyr Zelenskyy is a “dictator” with a public approval rate of 4%, and that Ukraine has started the continued conflict with Russia.
The controversial obvious optimization at Grok 3 comes than some criticize The model is to the left. This week, the users found that Grok 3 would consistently say that President Donald Trump and Musk deserve the death penalty. Xai patched the problem quickly; Igor Babuschkin, engineering director of the company, called It is a “really terrible and worse failure”.
When Musk announced GRK about two years ago, he provided the AI model as nervous, unsuccessful and anti-waxed-in general to answer controversial questions that other AI systems did not answer. He delivered part of this promise. GROK and GROK 2, for example, being vulgar would like to be happy to hire and spit out a colorful language that they would probably not hear about Chatt.
But Grok models in front of Grok 3 secured to political subjects and would not cross Certain limits. Actually, a study found that GROK leaned on topics such as transgender rights, diversity programs and inequality.
Musk has held the behavior of the training data of GROK – public websites – and pledged “Show closer to politically neutral”. Others, including Openai, be followedPossibly from the trump administration’s allegations for conservative censorship.
