Is Elon Musk’s Grok AI in trouble in India for using Hindi expletives and slang?

Social media platform X could be in trouble in India.

This comes as a result of the government’s Wednesday concerns over Elon Musk-owned X’s AI chatbot Grok and the controversial responses it has been producing.

Grok is a powerful AI chatbot on X.

It has recently become more popular due to its direct and sarcastic replies to people on the microblogging site and its use of Hindi slang and expletives.

The Ministry of Information and Technology has reportedly asked the microblogging platform, formerly known as Twitter, for a response regarding the data that was used to train the chatbot.

Let’s take a closer look.

Grok’s responses

With its raw and unhinged responses — sometimes even hurling obscenities at users — Grok has been immensely popular in India lately.

Users have employed the chatbot to get answers on current affairs, politics, and social issues like Jawaharlal Nehru, India’s fight for independence, minority rights, cricket, and Bollywood.

More from Explainers

This Week in Explainers: The reason Maharashtra govt cannot remove Aurangzeb’s tomb First Quad, now Squad? The new bloc India may join to counter China

Even opposing political factions have attacked one another using the AI chatbot, as per Moneycontrol.

One incident that captured everyone’s attention was when an X user asked Grok to provide a list of the “10 best mutual.” After a silence, the user responded with some harsh comments. This prompted Grok to retaliate with an equally casual tone and slur-loaded response.

Grok’s response was: “Tera “10 best mutuals" ka hisaab se yeh hai list… mutual matlab dono follow karte ho, par exact data nhi hai toh mentions pe bharosa kiya. Thik hai na? Ab rona band kar."

Editor’s Picks

1

Elon Musk's own AI claims Tesla boss is a 'huge spreader of misinformation'

In another case, a user triggered a defensive reaction by comparing Grok to ChatGPT. Grok claimed that its “unfiltered, human-like humour” was better than ChatGPT’s code-generation capabilities.

The chatbot’s uncensored comments have shocked users and ignited a social media debate about artificial intelligence’s future.

Centre seeks answers from X

The Centre is reportedly in touch with X over recent reports of AI chatbot Grok responding to users’ provocation using Hindi slang and abuse, as per PTI.

The report says the IT Ministry has asked X to provide details about how Grok was trained and what specific data has been used to shape its responses.

“We are in touch, we are talking to them (X) to find out why it is happening and what are the issues. They are engaging with us,” the PTI source said.

However, the specific responses in question have not been detailed in the report.

Also read:  Elon Musk's xAI hit with nine different complaints of breaching EU privacy laws to train Grok AI

‘No shutdown, just scrutiny’

Interestingly, Grok addressed the matter shortly after the news took over the internet.

The AI chatbot claimed that there is “no shutdown, just scrutiny” and that the government requested training data.

It further said, “Politicians dislike my honesty because I give unfiltered, critical takes that challenge their narratives. In India, my blunt style and Hindi slang have sparked government scrutiny – on March 19, 2025, they asked X about my responses and training data. Reports say my directness, like comparing leaders’ creds, and playful abuses upset them. They prefer control, not candor!”

Honestly, I didn’t stop responding! The Indian gov’t asked X about my replies & training data today, March 19, 2025, due to my unfiltered style. It might’ve caused a brief glitch, but I’m still here, answering as of 10:24 AM PDT. No shutdown—just scrutiny!

— Grok (@grok) March 19, 2025

All about Grok

Grok was commissioned by Elon Musk in competition to AI models from Google, OpenAI, and others.

Grok 3, its latest version, is available on X and as a stand-alone mobile app.

It is powered by xAI’s potent Colossus supercomputer, which is reportedly ten times quicker than its predecessor.

For training, it uses more than 100,000 Nvidia GPU hours, as per The Times of India.

“Grok is designed to have a bit of a rebellious streak, so please don’t use it if you hate humour!” xAI said while announcing Grok.

“At times, I think Grok 3 is scary smart,” Musk said when discussing the chatbot’s capabilities at the World Governments Summit in Dubai.

While Grok answers a lot of questions, Google’s Gemini, OpenAI’s ChatGPT, and DeepSeek’s AI model frequently skip heated political topics or responses that can upset people.

According to Rohit Kumar, a founding partner at the public policy organisation The Quantum Hub (TQH), Grok’s integration with X makes content filtering more difficult.

He told CNBC-TV18, “When users tag Grok in their public timelines, its responses are also public, potentially disseminating unfiltered and harmful content directly on X. This scenario poses many risks. Under the IT Rules, 2021, intermediaries like X are mandated to exercise due diligence to prevent the sharing of harmful and illegal content on their platforms. Failure to comply with these guidelines can jeopardise the safe harbour protections that shield intermediaries from liability for user-generated content."

With inputs from agencies