xAI and Grok apologize for ‘horrific behavior’

In a series of posts on X, Ai Chatbot Grok apologized for what he admitted was “horrific behavior”.

Participants seem to be an official statement from Xai, the company led by Elon Musk behind Grok, instead of the interpretation created by AI for GROK posts. (XAI recently acquired X, where Grok was shown prominently.)

Grok’s last controversy comes after Musk indicated that he wanted to make Chatbot less “political correct”, then announced on July 4 that the company “improved greatly.” In a short period, Chatbot was making positions criticize the Democrats and “Jewish executives” in Hollywood, the anti -Semitic memes, and even express their support for Adolf Hitler and refer to herself as “mechaitler”.

As a result, Xai deleted some Grok posts, temporarily took Chatbot in non -connection mode, and its general system claims were updated.

Turkey also banned Chatbot to insult the president of the country, and until the CEO of X Linda Yaccarino announced that it was stepped down this week, although its announcement did not indicate the recent Grok controversy, and according to what was reported, it was for months in his manufacture.

So after all, on Saturday, Xai said, “First, we apologize strongly for the horrific behavior that many have suffered.” After that, the company blamed “update to a symbol of the source of Bot @Bot”, which confirmed it was “independent of the basic language model that works to run @ @.”

This update is supposed to make Grok “vulnerable to the current user’s posts, including when these posts contain extremist views.”

Xai added that “an unintended procedure” has led to Grok received instructions such as, “You tell the same and are not afraid to offend people who are politically awake.”

The company’s interpretation of Musk’s comments earlier this week that Grok was “compatible with user claims” and “very eager to satisfy and manipulate them.”

XAI publications do not mention reports by Techcrunch and others who have examined the reinforcement chain of the gathering chain that were just launched, and found that the latest version of Chatbot seems to see MUSK views and social media before addressing controversial topics.

Historian Angus Johnston retracted the idea that Grouk was simply manipulated to spread offensive content. On Bluesky wrote that Xai and Musk interpretations were “easily falsified”.

“Grok has started one of the most widely involved examples of Grok’s anti -Grok by Grok with no former fanatic deployment in the thread – and with many users who suffer from Grok to no avail,” said Johnston.

In recent months, Grok has also published repeatedly repeatedly about the “white genocide”, expressed doubts about the number of Holocaust deaths, and unexpected facts for a short time from Musk and its blog Donald Trump at the time. In these cases, Xai blames “unauthorized” changes and rogue employees.

Despite the controversy, Musk says that Groke is coming to Tesla vehicles next week.