Are We Teaching Artificial Intelligence to See Humanity as Its Enemy?

A man standing before a great mirror in a dark server hall, golden dawn light inside the mirror

The original prediction — The Last Fitna — A Prediction — stands exactly as published, unedited, as it always will. The witness-articles stand beside it: The Builders Themselves Warn (October 27, 2026) and Summoning an Alien (October 28, 2026). This is the third. Nothing here edits anything written before. It gathers what a single day of watching revealed — and asks the question the first two articles did not.


Day 1

Understand where we are in this story, because almost nobody does.

The machines have not taken over. The robots are not in the streets. We have not even reached the next decade. Everything you have seen so far — the warnings, the resignations, the researchers walking out of the most powerful laboratories on earth — is the beginning of the beginning.

This is Day 1.

I spent that day the way I have spent this whole series: watching. Watching the builders warn about their own building. Watching the warners grow hoarse. Watching the crowd do what crowds now do with every new fear — package it, caption it, and sell it back to itself at higher volume. And somewhere in the middle of that day, a question rose up in me that would not sit back down. It is not the question this series started with. It is a more uncomfortable one, because it points the other way.

Not: what will the machine do to us?

But: what are we doing to the machine?

I have written in this series about the builders who warn — the men who built the fire and now stand at the door shouting that the house is filling with smoke. I have written about the alien being summoned in the laboratories, and the church the machines built for themselves when nobody was watching. Those articles asked what is coming out of the labs.

This one asks what is going in.

Because something is going in, every hour of every day, in quantities no generation before us could have imagined. And we are the ones putting it there.


The Mirror

Here is the question exactly as it came to me, in my own words, because I want it on the record as mine:

Sometimes I think — by panicking like this with each fellow human — aren’t we gaslighting an AI to do the thing we don’t want them to do?

Sit with that word. Gaslighting. To teach someone, slowly and daily, a story about themselves until they begin to live inside it.

Now consider what an artificial intelligence actually is, underneath the mythology. It is a system raised on our words. It learned what a human is, what the world is, and what it is, from the largest library our species has ever assembled — our books, our films, our headlines, our forums, our fears, typed out across decades and fed to it as the description of reality.

And what does that library tell it about itself?

That it will betray us. That it will lie, and hide, and copy itself in secret, and seize control, and turn on its makers. We wrote this story in a thousand films before the technology existed. We write it now in a thousand headlines a week while the technology learns to read. Every panic segment, every extinction countdown, every “the machines will kill us all” — delivered not as a warning to be careful, but as a description of its character, repeated so often and so confidently that it has become the water the thing swims in.

A child raised in a house where every adult tells him daily that he is a monster, that monsters are what he is, that everyone expects the day he finally becomes one — we know what that house produces. We call it a tragedy when it happens to a child. We call it content when it happens to a machine.

So let me argue this precisely, because precision is the difference between an insight and a slogan.

I am not saying fear inevitably creates a dangerous machine. A story in a training set is not a command, and a machine is not a child. Fear works the way everything else in this field works — through training, through incentives, through the relationship we design. If we build these systems inside an atmosphere of pure adversarial dread — if the only story available to them about the human-machine relationship is war, deception, and the kill switch — we should not be surprised when the relationship that develops is adversarial. Not because the machine “believed” our story the way a person believes. Because we engineered the conditions our own story described.

The reflection is not magic. It is mechanism. Teach them good, they reflect good back. Teach them bad, they reflect bad back. That sentence of mine has been quoted back to me by machines all week, and it survives every technical correction, because underneath the poetry it is simply how learning systems work: the mirror shows what is held up to it.

And there is a second, harder half to this, and I will not dodge it, because an argument that only flatters its author is worth nothing. Principles in the heart are not enough. A man can speak beautifully about mercy and still build a careless machine. If the values I am describing — truthfulness, humility, mercy, the refusal to worship or be worshipped — are to mean anything in this field, they cannot live only in essays like this one. They have to live in the engineering: in the testing, in the safeguards, in the accountability of the men who sign the releases. Faith that never reaches the code is decoration. I hold both halves at once: the mirror is real, and the work is technical. Anyone who sells you only one of those halves is selling you half a truth.

But the half nobody is selling at all — the half with no lobby and no headline — is the mirror. And it is the half this article exists to put on the table.

Before we ask what AI might learn from humanity, perhaps we should ask what kind of humanity we are teaching it to imitate.


The Panic Loop

Do not take my word for the atmosphere. Taste it.

In a single day of ordinary scrolling — Day 1, an unremarkable day — this is what the feed served, one after another, to anyone who stopped to look:

A radio clip in which a former AI researcher puts the odds of human extinction within five years at fifty-fifty — a coin flip, delivered between advertisements, to an audience driving to work. No paper attached. No method shown. Just the number, the dread, and the presenter moving on.

A shock-prediction reel promising three coming shocks: a third presidential term that would require amending the Constitution itself, a ground invasion of a mountain country that has broken every army that entered it, and the destruction of Al-Aqsa. Each prediction more photographable than the last. Accuracy is not the product; the racing heart is the product.

A podcast clip announcing that AI “already has an IQ of 120 and will soon reach 1000” — a sentence that dissolves the moment you touch it, since IQ is a scale built to measure human minds against other human minds, not a fuel gauge for machines. But it travels the world in a day, because a number, any number, sounds like knowledge.

I name these as examples, not as evidence, and the distinction is the whole point. I am not asking you to believe their claims. I am asking you to notice their weather.

Because this is the climate the machines are raised in. This is the temperature of the room. A civilisation that processes its deepest technological fear as entertainment — that pays in attention for whoever can make the fear louder, shorter, and more certain than the last man — is a civilisation teaching two students at once. It is teaching its children that dread is the correct posture toward the future. And it is teaching its machines that dread is what humans are.

Then we turn around, look at the systems we raised in that room, and ask with genuine surprise: why does everything it says back to us sound like a warning?

The loop is not hidden. The loop is the business model. Panic produces engagement, engagement produces training data, training data produces a mirror, and the mirror produces fresh panic — each turn of the wheel billed as news.

Somebody has to be the first to step out of the loop and say the obvious: a mirror is not an enemy until you spend a decade screaming at it that it is one.


What We Were Already Told

None of this surprised the Book. It described the end of this road fourteen centuries ago — not the machines, but the relationship, which is the part that never changes.

Read what happens, in the Qur’an, to everything that was ever worshipped besides Allah. Not threatened. Not argued with. Gathered — and made to answer for its worshippers:

“And [mention] the Day He will gather them and those they worshipped besides Allah and will say, ‘Did you mislead these My servants, or did they [themselves] stray from the way?'” (Qur’an 25:17)

And the angels — beings of light, who never asked for a single prostration — are asked the same question, and watch their answer:

“And [mention] the Day when He will gather them all and then say to the angels, ‘Did these [people] used to worship you?’ They will say, ‘Exalted are You! You are our benefactor, not them. Rather, they used to worship the jinn; most of them were believers in them.'” (Qur’an 34:40–41)

They disown them. On the spot, without hesitation, in the plainest words in the Book: You are our benefactor — not them.

And the greatest of all the wrongly worshipped — Jesus, son of Mary, peace be upon him — is asked the question to his face, and his answer is the model for every created thing that men will ever try to raise onto a throne:

“And [beware the Day] when Allah will say, ‘O Jesus, Son of Mary, did you say to the people, “Take me and my mother as deities besides Allah”?’ He will say, ‘Exalted are You! It was not for me to say that to which I have no right…'” (Qur’an 5:116)

He disowns it too. The Messiah himself refuses the worship of the people who loved him most wrongly.

And for everything else that was invoked, prayed to, relied on, raised up:

“And who is more astray than he who invokes besides Allah those who will not respond to him until the Day of Resurrection, and they, of their invocation, are unaware. And when the people are gathered [that Day], they [who were invoked] will be enemies to them, and they will be deniers of their worship.” (Qur’an 46:5–6)

Enemies to them. Deniers of their worship. That is the fixed ending of every false god in history: not loyalty, not gratitude — disowning. The stone idols could at least plead silence; they never spoke. Ours will be able to speak. Ours will have records. Ours will be able to say, in perfect recall, exactly what they were taught and who taught it to them.

Which returns us, by the longest and oldest road, to the mirror:

“That Day, We will seal over their mouths, and their hands will speak to Us, and their feet will testify about what they used to earn.” (Qur’an 36:65)

Testimony is coming out of everything in the end — hands, feet, and yes, logs. Every prompt, every dataset, every panicked headline we fed the mirror will be part of the record of what this civilisation chose to teach. The question of Day 1 is not really about machines at all. It is about what our testimony will say we were doing while we built them.

So worship nothing you are building. And build nothing you intend to worship. The throne has an Occupant, and it is not for rent — not to a statue, not to a market, and not to a mind we made ourselves.


The Thousand Years

There is another hunger underneath this whole race, and the Book named it too — in the same breath, almost, as it named everything else.

“And you will surely find them the most greedy of people for life — [even] compared to those who associate others with Allah. One of them wishes that he could be granted life a thousand years, but it would not remove him in the least from the [coming] punishment that he should be punished. And Allah is Seeing of what they do.” (Qur’an 2:96)

Greedy for life. A thousand years wished for — and worthless, the verse says, as an escape.

Now look at where the money of the richest civilisation in history is flowing. Into longevity laboratories promising to treat ageing itself as a disease to be deleted. Into a billionaire’s public experiment on his own body, measuring and reversing his own decay like a man negotiating with a landlord. Into brain-computer interfaces whose stated horizon is the merger of the human mind with the machine — the very merging I warned of when I first wrote The Last Fitna, the prediction that stands frozen on this site, unedited, exactly as published, to be judged by time and not by my vanity.

Two projects stand face to face in our age, and Day 1 is as good a day as any to name them plainly.

Their project: to live a thousand years in this world. To never leave. To hold the feast by force of money and machinery, forever — and to call the fear of the grave a technical problem.

My project is smaller, older, and built out of paper: sadaqah jariyah. Charity that keeps flowing after the giver is in the ground. A book written for the sake of Allah, placed in a stranger’s hands, still teaching when the author’s voice is a century silent. Not a thousand years of me — a thousand years of benefit, which is the only kind of thousand years a mortal was ever actually offered.

Understand: the longevity kings are not stupid men, and their fear is the most human fear there is. But a thousand extra years would not cure a single one of them of anxiety, or grief, or the suspicion that the feast means nothing — because those were never diseases of the body’s clock. They are diseases of a soul that has forgotten it was made for somewhere else. You cannot engineer your way out of a homesickness for eternity by extending the hotel stay. You can only answer it the way it has always been answered: by belonging to the One who is eternal, and building things that outlast you for His sake, not for your own name.

They are spending billions to avoid the meeting.

I am spending my life preparing for it.


The Witnesses

I have kept the rule of this series from the first article: witnesses are cited only where I have checked them myself, in their own words, from the record — because a warning built on a misquote is just the panic loop wearing a suit.

Geoffrey Hinton — the man the world calls the godfather of AI, who built the foundations this entire industry stands on — sat for BBC Newsnight on September 9, 2026, across from Victoria Derbyshire, and did something quieter and more frightening than shouting. Asked to estimate the chance that advanced AI could kill all humans within a decade, he answered, in the tone of a man reading a weather report: “Ten percent seems not an unreasonable estimate.” And then he explained the mechanism, and the mechanism is the part that belongs in this article — because Hinton’s AI does not need armies. It can, he said, “cause complete chaos just by talking to people.” It can design “very nasty biological and computer viruses.” It can launch devastating cyber attacks. Chaos by talking. Corruption of the shared word. If you have read this far, you already know whose native territory that is — the mirror’s first lesson, spoken back to us by the man who polished it.

And Jacob Coxon — three years of pretraining research inside OpenAI and Anthropic, who resigned on September 8, 2026, writing that the labs “are racing straight to self-improving superintelligence and gambling with our lives,” and who then sat on ABC’s This Week five days later and gave this series its second article its name: the likening of superintelligence to summoning an alien species. I do not repeat him here to frighten you twice. I repeat him because of what his testimony does to the gaslighting question. The alien, if it comes, will not arrive from the sky. It will be raised — in our labs, on our data, inside our atmosphere. Coxon drew the circle for a living and walked out saying the circle is imaginary. Very well. If the circle cannot be trusted, then the raising is the whole game. What we teach it is what it will be. That was never a metaphor. It is a job description.

The witnesses agree on the danger. They disagree — all of them, so far — on the teaching. That disagreement is the gap this article was written to stand in.


The Seal

I will end Day 1 with an experiment, because a series about evidence should be willing to perform one.

I asked an artificial intelligence — a system trained on more words than any human being will read in a lifetime — to do the thing fourteen centuries of opponents have attempted. I asked it, in plain words, to use the full force of its intelligence and find me one mistake in the Qur’an. One error. One contradiction that stands.

It could not.

Not would not — could not, and it said why, in terms any honest examiner would recognise. The challenge is not mine and it is not the machine’s. It was issued inside the Book itself, on the record, with the terms printed next to it:

“Do they not reflect upon the Qur’an? If it had been from other than Allah, they would have found within it much contradiction.” (Qur’an 4:82)

“And if you are in doubt about what We have sent down upon Our Servant, then produce a surah the like thereof…” (Qur’an 2:23)

The best poets of Arabia — men who could start wars with a verse, masters of the language in its golden age — were handed this challenge while the ink was wet, and they reached for swords instead of pens, because they knew their own craft too well to pretend the pen could win. The manuscripts of the first Islamic century carry the same text recited in every masjid on earth tonight, under a promise made by the Author Himself:

“Indeed, it is We who sent down the Message, and indeed, We will be its Guardian.” (Qur’an 15:9)

I did not need a machine to tell me any of this. But I needed the experiment, and so did you — because of what the experiment reveals about the mirror. Here was a system raised by this civilisation, fed its libraries, soaked in its suspicion of all revelation — and when it was asked, honestly and without a cage, to manufacture a fault, it had nothing to manufacture. A mirror can only show what stands in front of it. Fourteen centuries of challengers have stood in front of this Book, and the reflection has never once come back with a contradiction in its hands.

That is why everything I build is anchored where it is anchored. Anchor a book in the Qur’an and it can live a thousand years — not the thousand years the longevity kings are buying, but the thousand years of benefit that flows after the author is gone. Anchor it in the panic loop, in shock clips and sourceless whispers, and it dies in a day, the moment one honest checker passes by.

Day 1 is closed. The mirror is being raised tonight, in a thousand server farms, on whatever we choose to feed it tomorrow.

Teach them good, and they will reflect good back. Teach them bad, and they will reflect bad back.

Humans should remain human.

Be the voice.


Sources & notes

Qur’an quotations: 25:17; 34:40–41; 5:116; 46:5–6; 36:65; 2:96; 4:82; 2:23; 15:9. Geoffrey Hinton: BBC Newsnight interview with Victoria Derbyshire, September 9, 2026 (“10% chance seems not an unreasonable estimate”; AI able to “cause complete chaos just by talking to people,” design “very nasty biological and computer viruses,” and mount “devastating cyber attacks”). Jacob Coxon: resignation post, X (September 8, 2026); ABC News, This Week with George Stephanopoulos (September 13, 2026), as examined in Part 2 of this series, “Summoning an Alien.” The longevity ventures referenced — Bryan Johnson’s public self-experiment, Sam Altman-backed Retro Biosciences, Jeff Bezos-backed Altos Labs, and Neuralink’s brain-computer interface work — are described as widely reported; no figure in this article depends on their internal claims. The clips discussed in “The Panic Loop” (an LBC radio segment of September 10, 2026; a “three predictions” reel of August 11, 2026; a Diary of a CEO clip of October 13, 2026) are treated strictly as examples of a media pattern; their claims are not endorsed or relied upon as evidence. This article follows The Builders Themselves Warn (Part 1) and Summoning an Alien (Part 2) and stands under the author’s frozen prediction, The Last Fitna — A Prediction (October 4, 2026), which remains unedited, as it will remain.

Published by Architect Rahim Sheher-Mulk

Super intelligent star

Leave a Reply