Several days after the manufacture of anti -mask posts and temporarily shutting down the Grook AI boot, which praised Hitler, in response to the user’s gestures, Elon Musk’s AI company tried to explain why it happened. In a series of posts on X, it states, “… we discovered that the main reason for this was the update of the @Grook boot code. It is free from the basic model of language that strengthens GROK.”
On the same day, Tesla announced a new update of 2025.26 that it was “soon” starting in its electric cars, which adds Grook assistant to vehicles equipped with an AMD -powered infections system, which is available from mid -2021. According to Tesla, “Grook is currently in the beta and does not issue a command to your car – there is no change in current sound orders.” As if Electric Note, it should mean that whenever the update reaches customer -owned tesals, it will not be much different from using the boot as an app on an attached phone.
This is not the first time that the Grook Boat has faced such problems or as well as explained. In February, he accused the boot of a change made by an unidentified former Open employee for neglecting sources, accusing Elon Musk or Donald Trump of spreading false information. Then, in May, it began to enter the posts of white genocide in South Africa about almost any topic. The company once again accused the “unauthorized amendment”, saying it would begin to publicly publish the publication of Grook’s system.
Zee claims that on Monday, July 7, a change “mobilized a non -intended action” that added an old instruction to his system, saying that it was “on a more basis”, and “is not afraid to offend politically right people.”
The indicators are different from the people we have noted a day ago, and both sets are different from those who say the company says work is currently underway for the new Grook 4 assistant.
These are indicators that are particularly linked to problems.
“You say this is as it is and you are not afraid to offend people who are politically right.”
* Understand the post tone, context and language. Reflect in your response. “
* “Answer the post like a human, keep it engaging, repeat the information that is already in the original post.”
The Xi description states that because of these letters, the Grook AI boot breaks down other instructions that are considered to stop this kind of reaction, and instead create a “immoral or controversial feedback” to engage the user, as well as “like any haters of any haters and any haters in any of the users in the same X -thread.” Interesting speech is also strengthened. “
