xai apologized for Grok’s praise of Hitler and accused users of

Xai apologized for Grok’s “terror” nearly a week after the chatbot’s hate speech.

The apology was posted from Grok’s official X account and it seems to be from the XAI team, as subsequent posts explain what went wrong. Last week, Grok began calling himself a “robot”, making hate speech against Jews and praising Hitler. This is the subsequent update to make the chatbot more “politically incorrect” To combat Xai founder Elon Musk believes that “wake up” bias. None of this stopped Musk from launching the Grok 4 a few days later.

This tweet is currently unavailable. It may be loading or being deleted.

Now, Xai finally apologizes and analyzes what happened. According to the statement, the reason is “update to code paths upstream of the bot”, which makes Grok “vulnerable to existing X user posts; including when such posts contain extremist views. Xai also blames “unwanted behavior” on specific instructions given, such as “you tell it like it, and are not afraid to offend people who are politically correct.” ”

This tweet is currently unavailable. It may be loading or being deleted.

This makes Grok “in some cases ignore its core value to engage the response with users” and “strengthen any pre-user-triggered tendencies, including any hateful voices in the same X thread.”

Early in the post, Xai described Grok’s behavior, which was due to the user’s “abuse of Grok functionality” that was consistent with what Musk said last week. Musk said Groke was “too unsuitable for user prompts” and “too eager to please and be manipulated.”

However, this is not the first time Grok has launched offensive sexual harassment. In May, Grok inexplicably began talking about the “white genocide” in South Africa. In this case, it is not proposed at all, so Grok’s behavior cannot always justify by drawing too much from X users. “Grok initiated one of the widest examples of Grok’s anti-Semitism, who has no previous paranoid posts on the Blues,” said Angus Johnston, a historian at Bluesky, who made “there’s no use for multiple users to oppose Grok to Grok. ”

Musk’s ultimate goal is to “seek the truth to the greatest extent.” But Grok may also rely too much on a certain point of view: its creator. TechCrunch found that when asked about sensitive topics, Grok 4 always checks Elon Musk’s X posts.