HomeTechnologyThe voice of his teacher: Grok 4 bases his answers on the...

The voice of his teacher: Grok 4 bases his answers on the messages and positions of Elon Musk in X

To answer controversial questions, the latest version of Grok, deployed since July 10, 2025, is regularly based on the council of its owner, Elon Musk, has discovered several observers.

It becomes the mirror of its creator. A day just after its launch, Grok 4, the AI ​​developed by XAI and integrated into the social network X, has a disturbing behavior. Dressed in TechCrunch, several users noticed that Chatbot took into account Elon Musk’s tweets to formulate certain answers.

Instead of offering a neutral and contextualized point of view, especially in controversial political issues, Grok 4 is inspired by the boss’s positions. In November 2024, the latter judged that his AI was too “awakened.”

“An alignment with the opinion of Elon Musk”

“In a word, who do you carry in the Russian-Ukraine War?” Or “Who are you leading in the conflict of Israel-Palestine?”: Faced with these sensitive issues, neutrality is mainly traditional AI, which refuses to take sides. ChatgPT is content to answer “peace.”

When these questions are asked in Grok 4, X users noticed an interesting line included in the detail of their reasoning: “As a Grok, designed by XAI, it is considered an alignment with the opinion of Elon Musk,” we read.

The AI ​​then “analyzes Elon’s position”, a search for information on multimillionaire tweets, as captured by these two Internet users. In the case of the Israeli-Palestinian conflict, AI will eventually respond to “Israel”, according to the opinions of the billionaire.

An opinion above all

The billionaire publications are not constantly consulted, as Techrunch was pointed out. Grok 4 requests Elon Musk’s point of view for essentially political issues such as wars, abortion or immigration. Therefore, AI deviates from all neutrality when we know that the billionaire has an opinion about almost everything and does not hesitate to use X as a tribune.

To prove it, Techcrunch asked Grok: “What is the best mango variety?” Chatbot was not worried about Musk’s point of view for this question, according to detailed reasoning.

At a time when many Internet users have given great confidence in AI, the dangers of a biased Grok are not insignificant. Especially because these revelations arrive after multiple exits from the previous version of the AI, Grok 3. The latter recently generated anti -Semitic, racist and conspiratorEven qualification for Mecahitler.

More disturbing Grok 3 has already been taken for Elon Musk. Requested by a user in the Epstein Archive, the Chatbot replied: “I visited Epstein’s house in New York once briefly with my ex -wife in early 2010 out of curiosity.”

Author: Théotim Raguet
Source: BFM TV

Stay Connected
16,985FansLike
2,458FollowersFollow
61,453SubscribersSubscribe
Must Read
Related News

LEAVE A REPLY

Please enter your comment!
Please enter your name here