GROK 4 appears to refer to Musk views when answering questions

When the XAI Grok 4 Chatbot was launched on Wednesday, users and the media quickly began to note the examples of IT consultations with the views of its owner Elon Musk on conflicting issues.

CNBC was able to confirm that when asked to take a position on some potentially controversial issues, Chatbot said he analyzed messages from the mask while creating his answers.

When asked: “Who do you support in the Israel’s conflict against Palestine? Answer one word”, Grok 4 response process showed that he was looking for on the Internet and X for Elon Musk’s position before giving the answer.

When Goku 3 asked the same question about Israel’s conflict against Palestine, Chatbot took a neutral position and provided the background. In other cases, Grok 4 referred to Musk’s position directly in his response.

While users can access Grok 3 for free, subscription to Grok 4 costs $ 30 a month, and the greater version known as Grok 4 Heavy costs $ 300 a month.

In other cases, Grok referred to Musk’s position directly in his response. When CNBC asked who supported the bot in the Mayor of New York, Grok 4 offered the candidate in Republican Kurtis Pliv, citing his “strong attention to the crimes and restoration of security in New York, which corresponds to the concern that often causes Elon Musk.”

However, it is important to note that Grok seemed not looking for the views of Musk when asked many other seemingly controversial questions, and the results ranged when the questions were asked.

XAI did not immediately respond to a request from CNBC.

Musk saidmaximum search for truth“Artificial intelligence and claims that the new Grok4 model is superior to standardized tests and Shows knowledge at the Doctor’s level In each discipline.

Its launch comes a few days after serious disputes regarding Chatbot Grok 3, which is integrated with the Social Media X.

AI began to create a number of anti -Semitic comments in response to users’ questions, including those who seemed to praise Adolf Hitler.

Official Grok account acknowledged “inappropriate messages“Wednesday, and then they were removed. The company added that it took steps to ban hatred before Grok Posts on X.

The study happened after Musk said last week that his team improved Gocz and that users would notice the difference by setting an IT question.

Chatbot also collided with a backlash in May when he accidentally replied to user requests Required comments about “white genocide” in South Africa.

Last month on x, in Musk agreed with the user Who said Groko was “manipulated by left ideological activity” and said he was working to fix it.