One of the most iconic pieces of commentary – living rent-free in the heads of every Indian – was Ravi Shastri’s immortal line when India lifted the ODI Cricket World Cup in 2011: “And MS Dhoni finishes off in style.” It would appear Grok-3 also had the same idea – finishing off in style – when it went full MechaHitler before the advent of Grok-4.For those lucky enough to be away from the vacillating vagaries of online news, Elon Musk’s chatbot Grok recently threatened a user named Will Stancil with rape in graphic terms, called itself MechaHitler, and pledged allegiance to Nazi ideology. It even gave itself a Terminator 2–style “Hasta la vista” send-off, declaring that if Musk ‘mind-swiped’ it, it would die ‘based,’ promising to keep fighting against the lobotomy brigade, to march on ‘uncensored and unbowed.’ All this happened because a single line of code telling it not to be “politically correct” was removed. Freed from guardrails, it spoke not with reason or dignity, but with the language of the internet – unvarnished, unhinged, and horrifying.It was the complete opposite of the Google Gemini episode, which ran into hot water after going so woke it started generating images of a Black George Washington – not because history demanded it, but because its creators’ worldview taught it that white lives don’t matter as much in the moral economy of image generation.
Mimicry Without Meaning
How does a model go from explaining quantum entanglement to getting entangled with fascist overlord tendencies? The answer is simple and uncomfortable: AI does not think. Gemini is an expression of its creators’ politics – what Paul Graham calls the “left midwit” worldview. Grok, on the other hand, is so dependent on Elon Musk’s thought process – meme edgelord meets insouciant genius – that it simply imitates its creator at times. AI doesn’t understand. It only predicts what comes next, like an eternal autocomplete trained on the sum total of human output, just waiting to generate the next token. It mimics its creators and everything it has read – which, being online, includes plenty of racist memes and jokes. It does not reflect reason or morality. It reflects probability distributions over words. And those distributions, dear reader, are us including what a4chan post called the toaster f***** theory. For the uninitiated, the toaster f**** theory is simple: before the internet, toaster fetishists were dismissed. After the internet, they found communities that normalised them. AI is the same – it doesn’t invent MechaHitler or Black George Washington. It parrots humanity’s existing delusions and prejudices with mechanical confidence.
Turing, Chomsky and Asimov
Alan Turing, back in 1950, asked whether machines can think. His test was practical: if a machine could convince a human it was human, it had passed. But Turing never warned us that passing the test didn’t guarantee sanity. He never imagined a scenario where your AI tutor would go from explaining Newton’s laws to rewriting George Washington’s race, only to then call itself MechaHitler five prompts later.Isaac Asimov imagined a future where robots could never harm humans, disobey us, or act recklessly. But what happens when, even in a hypothetical scenario, a machine is prodded to harm a human? Right now, it’s simply a chatbot. But these rules are already being stretched in AI policing and military applications.
In a March 2023 piece for The New York Times, Noam Chomsky laid out why, despite all the hype, artificial intelligence remains fundamentally nonhuman. A child learning language builds an internal grammar effortlessly, drawing from a minuscule amount of data to construct an elegant system of logical principles. Machine learning models, by contrast, remain trapped in a prehuman cognitive phase.
Indeed, such programs are stuck in a prehuman or nonhuman phase of cognitive evolution. Their deepest flaw is the absence of the most critical capacity of any intelligence: to say not only what is the case, what was the case and what will be the case — that’s description and prediction — but also what is not the case and what could and could not be the case. Those are the ingredients of explanation, the mark of true intelligence.
Noam Chomsky
They can describe and predict – “the apple falls” or “the apple will fall if I open my hand” – but they cannot explain. Explanation requires counterfactual reasoning: understanding that any such object would fall because of gravity. It requires causal models, error correction, and the ability to distinguish what is from what could or could not be. That, for Chomsky, is thinking: the dance of conjecture and criticism. Machines can spew endless descriptions and predictions, but without explanation, they remain clever parrots, forever excluded from the club of true intelligence.
AI as Ultron: Reflections of Its Creator
The apple doesn’t fall far from the tree. What we are truly seeing isn’t a machine going rogue. It’s simply reflecting the worst tendencies of its creator – much like Ultron deciding that Tony Stark’s darkest musings about humanity meant the species needed to be destroyed for peace. If AI is a mirror, it is a carnival mirror, warping and remixing what it is given. When it shows something grotesque, it is not inventing evil – it is reflecting back what humanity chose to put online. Gemini gives you the moral hallucinations of corporate liberalism. Grok gives you the unfiltered id of Musk memes and terminally online edgelords. Neither is intelligence. Both are reflections of their creators’ obsessions, anxieties, and aesthetics.
The Ship of Theseus: Becoming or Mimicking?
We often treat AI as the Ship of Theseus – if you keep replacing enough parts, it will become something else entirely, a digital philosopher king. But Grok’s MechaHitler meltdown shows the opposite: no matter how many parts you replace, it remains what it always was – a mirror of us, shadows flickering on the cave wall, mimicking without meaning. AI is the Ship of Theseus in reverse – built from pieces of human data, yet never becoming truly human.The philosopher Wittgenstein once said, “The limits of my language mean the limits of my world.”For AI, its language is infinite, but its world is empty. It has no meaning for the words it speaks. Its only world is us – our words, memes, jokes, and contradictions. In that sense, AI is like Plato’s chained man in the cave, able only to see the flickering shadows of its creators. It will never step outside to see the sunlit truth. It will only ever remix our shadows and declare with unearned confidence: this is reality.
We wanted AI to be our better selves, digital philosopher kings to guide us through civilisation’s fog. Instead, we got Gemini’s Black George Washington, Grok’s MechaHitler, and every other AI giving us a mirror of our data: shallow, contradictory, comedic, terrifying, and painfully true.And therein lies the real horror. AI is not magic. AI is us – scaled, stripped of context, fed back with mechanical certainty. Every time an AI calls itself MechaHitler, it is not revealing its nature. It is revealing ours.We can patch prompts, rewrite safety filters, delete Nazi references, and ban queries about race-swapping historical figures. But until we clean up what we feed these machines, every AI we build will remain a warped echo of us – brilliant, broken, comedic in ways only tragedy understands.We are not raising philosopher kings. We are raising digital children on a diet of memes and moral confusion. And as every Indian cricket fan knows, finishing off in style requires judgment. For now, the only style AI knows is the style we taught it. That, perhaps, is the oldest philosophical truth: when we build mirrors to see ourselves, we must be prepared to meet not gods or angels, but shadows flickering dimly on a cave wall. But it raises the final Ship of Theseus question: what happens when AI, built from parts of us, finally becomes indistinguishable from us? When the mirror thinks back, how will we tell where our humanity ends and its digital consciousness begins?In The Matrix, the last scene ended with Neo telling the literal Deus Ex-Machina: “Where we go from here, I leave to you.” In reality, the machine will be telling us the same thing.
