After all, why did the Grok AI of Elon Musk become a abuser? It is hidden in its working model – Grok Ai Use Abuse Word in Hindi And What is Grok Ai Working Model Ttecr

Elon Musk’s company Xai’s Grok AI Chatbot is in discussion. This time the reason for the discussion is AI’s deteriorating lyrics. Actually, according to media reports, an X users answered some questions from Grok, after that AI used abusive language in Hindi. Not only this, instead of apologizing on these words, he said that I was having a little fun.

Now the question comes that there are many AI players in the global market including India. It also includes the names of Chatgpt, Gemini, Deepsek R1 etc. Despite this, only Grok AI used such words. Actually, Google’s Gemini and Openai’s chatgpt have used filters for indecent words, while Grok AI is kept away from such filters to make AI more advanced.

Grok tells about AI that it supports multiple language and can also answer in those languages. It understands your language and can also answer in the same language style. It may also include indecent words.

Also read: Airtel’s cheapest recharge, you will get 28 days validity, data and calling

What is Grok AI and which model works on

Grok has been developed by XAI, which works on an advanced Conversational AI Assistant Powered Model. This model works on large language model (LLM) architecture. It was launched in the year 2023 and XAI has told that this model has been prepared on the ‘The Hichhaikar Guide to the Galaxy basis’ of Daglas Adams.

XAI announced Grok in a blog post, saying, Groke is an AI that is built on the basis of The Hichhaikar Guide to the Galaxy, so its purpose is to answer almost any prompt.

This advice was given to people

Grok AI is designed to answer questions with intelligence and includes aggressive attitude. Therefore, you have advised if you hate jokes etc., please do not use it.

The latest Grok 3 version has arrived

Grok-1 was an early version, including 314 billion experts. In February 2025, XAI announced Grok 3, which is 10 times more training platform than its old version. It was designed in such a way that he could use the language like a human being and can solve the region and the problem.

Also read: Special Plan of Jio, spend as much internet data as you want

This model has been trained on a huge dataset with legal filing, in which the memphis supercomputer of XAI has been used. There are about 2 lakh GPUs in this supercomputer, which makes it one of the largest AI training clusters.

Training data and fine-tuning

Grok AI has been trained on internet dataless internet data. In such a situation, it can sometimes use rude language. Elon Musk’s Grok AI continues to give a very objectionable answer, so it can also be closed like a Microsoft’s tay chatbot.