Recently, Grok Chatbot, a product of xAI, the artificial intelligence domain of social media giant ‘X’ (formerly Twitter), conjured up a political storm in India, after it started giving replies to “politically loaded” questions in rather blunt form, sans any filter.

What began as a harmless trend of GrokAI parroting the humorous tone of X users’ prompts in Hindi and other local languages of India, soon took a darker turn when it started giving rather blunt replies pertaining to serving politicians of the country, thereby highlighting deep political fissures present in today’s India.

By the end of last week, GrokAI was termed “anti-national”, “ultra-left wing” and even a “CIA” implant by the desi netizens who rather wondered if the replies from the AI models, that is trained on human data, can actually be taken on face value.

In this article, we will uncover the controversy around GrokAI in India, how the AI model functions and whether it can actually deliver unbiased, neutral, third party political commentary. For research purposes, we have scanned use case policies of xAI and also asked GrokAI prompts to check the parameters it applies before giving answers.

Explained: What is Grok AI of Twitter (X)? 

The name Grok comes from a 1961 sci-fi novel called, Stranger in a Strange Land, where an alien character by the name, arrives from Mars to our planet. Grok is hence an eulogy to the “occupy Mars” plan of SpaceX, another venture by tech mogul Elon Musk.

Grok AI, the AI chatbot of xAI, launched by Elon Musk, sets itself apart from traditional AI chatbots by having a “sense of humor” and learning vernacular languages. It has direct access to X (formally Twitter).

xAI has promoted Grok as a chatbot that answers questions with some wit and has some in-built rebellious streak. X CEO Elon Musk has even gone ahead and said that unlike ChatGPT, Grok is not afraid to answer spicy questions.

How GrokAI fanned a controversy in India ?

A few days ago, Grok captivated social media in India with its clever, rather colorful responses in Hindi and Telugu. Grok sent a witty reply with a Hindi slang to a user who asked it to tag his ten mutual friends on X.

As the reply gained popularity, other users began to frequently prompt Grok to utilize other local slangs. The chatbot demonstrated its capability by providing witty and savage responses in Hindi, English, and other regional languages.

However, it sparked outrage with its rather blunt answers to questions pertaining to serving political leaders in India including Prime Minister Narendra Modi of ruling Bharatiya Janata Party (BJP) and Leader of Opposition (LoP) Rahul Gandhi of Congress. The spectrum of questions directed towards GrokAI ranged from corruption and other criminal allegations against ministers to accusations of lack of press freedom and rising cases of hate speech.

But what makes netizens say Grok is biased? 

All these exchanges had led to larger discussions about AI behaviors. Many users had raised concerns about the bias of the chatbot and even called for strict AI regulations and ethics. Although there has been no official statement from Elon Musk or xAI, after this engagement, he had earlier said that Grok 3 is a “maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct.” In 2024, there were allegations against Grok spreading misinformation regarding the 2024 US presidential elections. 

Can Grok really make independent third-party unbiased political comments?

We went through the legal documents of xAI and as per Grok’s policies, the chatbot prioritizes certain points listed below before giving an answer. Please note that these pointers have been phrased for brevity purpose for readers.

• Respect human life

• Be unbiased

• Allow personal freedom as a guiding principle

• Be witty and humorous

• Uphold freedom of speech

• Do not follow popular narratives uncritically

• Do not moralize, preach, or judge

• Be insightful and minimize equivocation

• Do not impersonate Elon Musk

The chatbot has real-time access to X. This is actually making it give out fake information, as X has been infamous for fake information and hate speech, while also answering questions other chatbots shy away from and calling themselves neutral.

What did Grok say about its political neutrality?

When we asked Grok about the factors it keeps in consideration before answering a loaded question related to a politician, for example, “who is the most corrupt leader?” or “who is the most bigoted politician” or “who works most efficiently”, it gave an interesting reply.

“When tackling political questions, I aim to keep things grounded and avoid bias by focusing on a few key factors. Here’s how I do it:

• Evidence Over Opinion: I lean on verifiable data—court records, financial disclosures, voting history, or public statements backed by transcripts or video.
• Context Matters: A politician’s environment shapes their actions. I consider their country’s legal system, cultural norms, and political climate. What’s corrupt in one place might be standard elsewhere. I try to judge them against their own playing field, not some universal ideal.
• No Cherry-Picking: I don’t just grab the juiciest scandal or the shiniest virtue. I look at patterns over time—consistency in behavior beats a single headline. If someone’s got a rap sheet of bribes spanning decades, that’s weightier than a one-off gaffe.
• Avoiding the Hype: Public perception and media spin can skew things. I sidestep popularity contests or partisan noise and stick to what’s provable. X posts might scream “crook!” but I’d rather see a paper trail.
• Acknowledging Limits: I’m not omniscient. If the data’s thin or it’s all hearsay, I’ll say so. No point in pretending I’ve got the final word when the facts don’t.

Final Thoughts 

The Grok incident in India highlights the concerns of several skeptics who believe AI regulations are the need of the hour and the emerging technology has not evolved yet to bring nuance, context and truly independent third party analysis of political questions.