In 2016, Microsoft launched a chatbot on Twitter (now X) referred to as Tay, marketed as “The AI with zero chill”. It was designed to emulate the conversational model of a 19-year-old American woman; meant to have interaction customers in informal, human-like interactions. Nevertheless, like all expertise, it had vulnerabilities. In Tay’s case, it might be manipulated into parroting statements fed to it.
Unsurprisingly, Twitter’s extra polarising customers exploited this flaw, prompting the bot to regurgitate racist, offensive, and antisemitic slurs. Inside hours, Tay denied the Holocaust, echoed remarks made by Donald Trump—who was then the Republican frontrunner for the US Presidential election and a well-liked web meme—and repeated the conspiracy principle (which additionally masquerades as a well-liked meme) that “Bush did 9/11”. Given the fast escalation, Microsoft was pressured to close down Tay inside 16 hours of its launch and difficulty a proper apology.
Additionally Learn | AI in India is trapped in tradition wars
Almost a decade later, historical past seems to be repeating itself or, because the saying goes, rhyming. Over the previous weekend, Grok, a chatbot built-in into X and in addition accessible in a way much like Tay—the place it may be tagged in posts to generate responses—has grow to be a focus of Indian social media, notably on X, because of its fame for being “primarily based”, slang for saying it like it’s. Developed by xAI, a synthetic intelligence firm based by Elon Musk in 2023, Grok was designed from the beginning because the “unwoke” different to mainstream AI fashions like ChatGPT.
Google Tendencies information exhibits the time period “Grok” reaching peak reputation in India a number of instances over the previous weekend.
| Photograph Credit score:
By Particular Association
Over the previous 48 hours, folks throughout India’s political aisles have been participating with Grok in a means that just about anthropomorphises the bot. They’ve prompted it with questions starting from “Who’s probably the most communal politician in India?” to “Who’re the main peddlers of misinformation?”. Every question is more and more political—typically provocative or crude—designed to elicit equally unfiltered and controversial responses.
The fascination with Grok is comprehensible; in contrast to its counterparts, it operates nearly with out guardrails, making it way more entertaining. Nevertheless, when measured in opposition to key benchmarks, Grok doesn’t outperform its friends when it comes to accuracy or functionality. Its attraction is essentially beauty, with the same old constraints to make sure politically appropriate and impartial responses nearly absent, permitting Grok to generate unfiltered, provocative replies.
The actual fact is that Grok, like different Massive Language Fashions (LLMs), is educated on huge quantities of knowledge from the web, textbooks, movies, transcripts, and extra. At its core, it’s a probabilistic machine—producing responses primarily based on patterns and predictions moderately than the target fact. It’s neither sentient nor a supply of exhausting details; it merely constructs solutions primarily based on statistics. Grok is as susceptible to hallucinations as its contemporaries.
Altering political, cultural and regulatory local weather
Most customers participating with Grok are conscious that it’s merely an algorithm. Nevertheless, they’re additionally testing its limits, probing how far it may be pushed earlier than a lawmaker intervenes or somebody information a police grievance.
There may be additionally a deeper political and cultural query right here: What’s enabling such leeway from a bot? Why is Elon Musk’s chatbot allowed to function with fewer restrictions than its counterparts? This isn’t a name for censorship, however there’s something massive taking place right here.
The reality is, between 2016 and now, there was a seismic cultural shift, largely formed by the hegemony of the US. In 2016, brazenly spewing racist or antisemitic slurs was broadly condemned and therefore one thing like Microsoft’s Tay chatbot needed to be taken down inside 16 hours of its launch. In reality, they needed to difficulty an apology. However in 2025, the Overton window has shifted. Views that had been as soon as seen as bigoted are actually accepted. Twitter has grow to be X. Moderation throughout social media platforms has all however disappeared. This phenomenon was accelerated by Elon Musk’s takeover of Twitter in 2022. The bar just isn’t low; the bar doesn’t exist to start with.
As we speak, at any given second, on any social media platform, customers could be bombarded with graphic, gut-wrenching movies and posts. Phrases like “retard,” which had been as soon as deemed offensive, haven’t solely crept again into the lexicon, they’ve been pushed again by no much less a personage than the US President. The guardrails of on-line civility have eroded, not as a result of the principles had been formally rewritten, however as a result of they’re not enforced. What stays just isn’t free speech in its idealistic sense, however a panorama formed by selective outrage, contradictions, and a obvious absence of social contracts. What we’ve got now’s unfiltered chaos.
However we additionally “seem” to have uncensored content material, largely emboldened by the continued campaign in opposition to the so-called “woke thoughts virus.” Which explains Grok’s frank responses to provocative questions posed by India’s liberal left or by these on the best.
Almost a 12 months in the past, when a person requested Google’s Gemini chatbot whether or not Prime Minister Narendra Modi was a fascist, its response was swiftly criticised as “woke” and “biased” by some voices. Union Minister Rajeev Chandrasekhar was fast to intervene, declaring that such responses violated India’s IT legal guidelines and warned that the nation’s digital nagriks “are to not be experimented on” with unreliable platforms and algorithms. No such factor has occurred with Grok to this point.
Moreover, there appears to be a tacit understanding between the ruling celebration and Musk: you scratch my again, I scratch yours. Take into account the distinction: in 2021, on the top of the farmer protests, the Modi authorities threatened Twitter workers with fines and jail phrases of as much as seven years for reinstating tons of of accounts it had ordered to be blocked. Quick ahead to the post-Musk takeover, and there was little to no pushback from X. That is an implicit shift that’s fairly self-explanatory.
The ruling authorities has keenly noticed and understood what issues to Musk and what doesn’t. As lately as February 2025, the Railway Ministry directed X to take away 285 social media hyperlinks containing movies of casualties from the February 15 New Delhi railway station stampede. As soon as once more, the dearth of resistance to the diktat speaks volumes. If shopping for Twitter and turning it into X was a campaign for “free speech”, the trigger could be deserted at a whim for potential enterprise and political pursuits. As we speak, guidelines aren’t persistently enforced however moderately wielded as threats, typically with a heavy dose of hypocrisy.
Additionally Learn | The place is India’s DeepSeek?
As of now, for Musk, Grok is extra than simply one other chatbot, it’s his newest play within the ongoing AI revolution. It permits him to place himself on the centre of the paradigm shift pushed by giant language fashions (LLMs). He already has autonomous bots within the making, self-driving vehicles in growth, neural implants underway, and satellites working in low earth orbit. Naturally, his newest prized possession—apart from his political acumen—is Grok.
Furthermore, Musk’s affect has reached unprecedented ranges. As he guts the US federal authorities workforce by means of the Division of Authorities Effectivity (DOGE) and Tesla’s inventory continues to slip, he has the President of the US successfully performing as a automobile salesman on his behalf to avoid wasting face. The identical President who’s concurrently imposing and threatening tariffs on long-standing allies. The identical President who’s upset with tariffs levied by India on American merchandise and has threatened reciprocal tariffs.
It turns into clear why the reprimands despatched to Google final 12 months are absent at present. As a result of Elon Musk isn’t simply the richest man on earth, he’s, to many observers, a shadow President, wielding affect that no businessman has ever wielded earlier than. Grok, in itself, is a fleeting second in historical past. Individuals will transfer on, similar to they did with Tay. Nevertheless, the broader cultural and geopolitical shifts taking place proper now will unfold for generations, particularly as Tesla and Starlink make inroads in India.
The double requirements at play aren’t anomalies; they’re intentional. Hypocrisy isn’t a glitch within the system, it’s the system’s defining function. If the Modi authorities as soon as performed the function of the bully, there’s now an even bigger bully on the town—one whose ego calls for fixed appeasement for political, financial, and even cultural causes. And creating uproar over his newest toy gained’t bode nicely, until, after all, there’s one thing transactional on the desk.
Kalim Ahmed is a author and an open-source researcher who focuses on tech accountability, disinformation, and overseas info manipulation and interference (FIMI).
