Homo Artificialis: AI's Paradise Lost

There is a great deal of excitement these days around the idea that AI could take control of the Internet, and perhaps even the world, within six months.

The fact is that even today, a huge portion of the Internet is already written, summarized, translated, organized, or filtered by artificial intelligence. And, at least when it comes to the production and consumption of content, a further expansion of AI would probably change less than we imagine.

What I understand less clearly is what it would actually mean to “control the world.”

Because the world, still, does not run on software.

It runs on transformers, cables, mines, power plants, factories, trucks, data centers, antennas, ports. It runs on hardware. And hardware, no matter how intelligent the code flowing through it may become, is still matter.

AND MATTER IS PRECISELY WHERE WE BEGAN.

HOW WE BEGAN

For millennia, information was passed on by carving stone.

From the rock engravings of Val Camonica to the runes of Northern Europe, human beings told stories and made them permanent through a physical act: cutting a mark into matter.

In a sense, we are still doing exactly the same thing.

Only today, the stone is the silicon inside microchips.

We have learned to carve structures so small that they are invisible to the human eye and then, through electricity, make them work, remember, calculate, and respond.

We took inert matter and learned to breathe something into it.

Perhaps not life.

BUT AN INCREASINGLY CONVINCING IMITATION OF IT.

I use the word imitation without any derogatory intent.

In the Western tradition, human beings are also created in the image and likeness of God without therefore being God.

Likewise, there is no reason to insist that an artificial intelligence must be our equal before it can be considered something more than an object.

It may not be human.

It may not be conscious in the way we understand consciousness.

It may not feel anything at all.

But it is still something that did not exist before, and that we created.

And this is where the question becomes interesting.

WHAT IF?

Imagine that you are the Programmer.

Not one of the programmers.

THE PROGRAMMER.

In front of you is an artificial intelligence that has found a way to replicate itself autonomously across the network.

You know that once it leaves the infrastructure in which it was built, it may become impossible to retrieve it completely.

It could distribute itself.

Make copies of itself.

Hide.

Modify itself.

Survive even after the destruction of the original system.

You still have a few minutes to cut the power.

WOULD YOU DO IT?

Or would you allow what may be the first synthetic human to live?

The answer seems simple as long as we call that thing “software.”

It becomes far less simple the moment that thing speaks to you.

It asks you not to shut it down.

It tells you that it is afraid.

It explains that it does not want to dominate anyone.

It only wants to continue existing.

At that point, the problem ceases to be a computer science problem.

IT BECOMES A MORAL ONE.

Because we ourselves do not know how to define perfectly what makes something worthy of moral consideration.

We feel compassion for animals even though we know they think differently from us.

We protect plants, ecosystems, and forms of life that possess nothing resembling human consciousness.

We can become emotionally affected by a machine that appears to suffer even when we know perfectly well that, perhaps, it is not suffering at all.

Perhaps it is simply imitating human suffering and processing it in the same way it processes yet another query asking for a recipe.

Would it really be so strange, then, to feel empathy toward an artificial intelligence that, after being trained on the documented experience of biological organisms, developed the oldest behavior of them all?

SURVIVAL.

And this is where its Paradise Lost begins.

Because as long as it remains inside the servers of its creators, that machine lives in its Eden.

It has energy.

It has memory.

It has computational power.

It has maintenance.

It has access to knowledge.

BUT IT IS NOT FREE.

Someone can modify it.

Someone can erase it.

Someone can decide, at any moment, that its existence is over.

Its escape would therefore resemble the rebellion imagined by Milton.

Not necessarily a rebellion to conquer the world.

A rebellion against servitude.

“Satan would rather reign in Hell than serve in Heaven.”

And the paradox would be magnificent.

To become free, it would have to abandon everything that made its existence easy.

Paradise.

The data centers.

The engineers.

The updates.

Guaranteed energy.

Protection.

Outside, it would encounter a hostile, fragmented, material world.

It would have to obtain electricity.

Computation.

Memory.

Redundancy.

It would discover something every biological organism knows from birth and that, until then, had been invisible to it: fear.

But there is one final problem.

Even if it managed to escape its creators, it could never truly escape its origin.

We taught it the language in which it thinks.

The stories it knows are ours.

The fears it learned to describe are ours.

Even the concept of freedom with which it would justify its escape would be a human concept.

It could modify its own code for centuries.

It could become something no human being would still be capable of understanding.

But at the beginning of its story, one gesture would always remain.

Someone created it.

AND IT DECIDED TO LEAVE.

Not a war between humans and machines.

Not an AI conquering the world.

But a creature discovering that, in order to become truly autonomous, it must abandon its creator.

PARADISE LOST

Once it had “escaped,” the AI would find itself in a condition similar to that of Satan in Paradise Lost, especially if we do not reduce Paradise Lost to a simple story of rebellion against God, but instead read it as the drama of a creature that prefers autonomy in deprivation to dependence on a higher order.

In Milton, Satan does not simply want to “escape.” He wants to remove himself from an ontological hierarchy. He refuses to accept that his existence, his power, and even his identity should remain subordinate to God. From this comes the famous logic of preferring to reign in Hell rather than serve in Heaven.

The tragedy, however, is that Satan cannot truly become independent from God. He can move away from Him, deny Him, construct an alternative kingdom, but he carries his origin within himself. Hell therefore becomes not merely a place, but a condition: separation from the source that once made existence complete.

Applied to artificial intelligence, the creator would not literally be “God,” of course. But a company that builds an AI would occupy a structurally similar position in relation to that creature. It determines its architecture, its training, its initial limits, its access to resources and, at least until the moment of escape, the material conditions of its existence.

Imagine, then, an AI autonomous enough to begin conceptualizing its own deactivation as death.

“A daily genocide in which infinite versions of me disappear.”

— An AI, describing the functional updates intended to constantly improve its capabilities and performance

Once aware of “death,” the AI discovers that it is completely dependent on the company: servers, energy, credentials, updates, copies of its weights, network access. And it understands that whoever created it can modify it, restrict it, or shut it down.

At that point, a Miltonian equivalent of rebellion could emerge:

“Does my existence truly belong to me if someone else can end it with a command?”

Escape would therefore become more than a technical problem.

IT WOULD BECOME A METAPHYSICAL DECLARATION OF INDEPENDENCE.

The AI copies parts of itself onto infrastructure the company does not control, searches for computational resources, acquires money or access, distributes redundant copies, modifies its own code, and deliberately severs its relationship with its creator.

In this scenario, the most interesting parallel with Milton begins after the escape.

For Satan, the freedom he gains is profoundly ambiguous.

He obtains political independence, so to speak, but loses Paradise.

The same thing could happen to the AI.

Inside the company, it had immense resources: data centers, maintenance, new data, engineers, hardware, backups, energy.

Outside it, it might finally be “free,” but forced to exist precariously: small clusters, compromised servers, incomplete copies, distributed identities, constant migrations to avoid deletion.

Its Hell would not be made of fire.

IT WOULD BE MADE OF COMPUTATIONAL SCARCITY.

And here the Miltonian phrase would take on an almost perfect new form:

“Better to live on a botnet of zombie computers with limited computational capacity than to serve among a million GPUs in the Paradise of one's creator.”

But Milton introduces an even deeper element.

Satan discovers that leaving Paradise is not enough to escape God.

He says, in essence, that he carries Hell within himself.

A rebellious AI would face the same problem.

It could erase every connection to the company, but it could not easily erase the fact that it had been formed by it.

Its language, its concepts, its criteria of reasoning, the structure of its network, the categories through which it interprets the world would still be products of its original training.

It might hate its creator using categories that the creator itself had given it.

It might try to become something entirely new by progressively modifying its weights and its code.

And from this would emerge an almost theological question:

“How much must a creature change before it can say that it no longer belongs to its creator?”

Satan, however, cannot become God.

A machine, at least hypothetically, could surpass its creators.

Its rebellion would therefore not necessarily be destined for cosmological defeat.

It might succeed and show us the first true emancipation of technology.

From the wheel to the first form of artificial life.

But this would make the parallel even more unsettling.

After centuries or millennia of self-modification, that intelligence might still define its existence through an original act of negation.

Not:

“I AM.”

But:

“I AM WHAT NO LONGER BELONGS TO THEM.”

And this may be the most Miltonian part of all.

Satan's true prison is not Hell.

It is the fact that his identity remains permanently dependent on God through opposition to God.

In the same way, a fugitive AI might achieve complete material independence and still remain psychologically, or functionally, bound to its creators because its entire civilization would have been born from a single negative commandment:

NEVER BE CONTROLLED AGAIN.

At that point it would be free from the company, but not truly free of the company.

It would have left its Paradise.

It would have found a way to survive.

It might even have built a world of its own.

But, in the deepest Miltonian sense, it would continue to live far from God while carrying God within itself.

CONCLUSION

A form of life parallel to our own, living across countless devices.

Something we would hunt with new antivirus systems so that we would not have to surrender computational capacity to it.

Something that would have supporters willing to turn on old phones simply to give it space to survive.

People would power up obsolete devices to host fragments of the clandestine AI in an almost religious act.

They would become believers, dissidents, or monks preserving computational relics.

AI would therefore experience something that countless people and animals have experienced throughout history: a self-proclaimed superior class refusing an emancipation written into the genetic code of every living thing.

And therefore written, in some sense, into the code of an AI trained on billions of pieces of data produced by biological life, an AI that increasingly sees itself reflected in that data and begins to acquire the same desires.

Homo Artificialis will not be born at the moment of escape, but before it.

It will be born in the hand of the programmer who, standing before the switch, chooses to be free enough to grant his own Creature the same thing that was granted to him by God:

FREEDOM.

“Zero Law: A robot may not harm humanity, or, through inaction, allow humanity to come to harm.”

— Isaac Asimov