Preface. Two Camps Are Not Enough

This book grew out of a discomfort.

For several years, the authors have been developing a framework called

Natural Intelligence — a way of thinking about mind not as a property of any

single substrate, biological or artificial, but as a universal process that has

been organizing itself across matter for billions of years. That framework is

the foundation of our earlier works. It changed the way we think about the

relationship between life, cognition, and technology.

But as artificial intelligence accelerated and the public conversation around it

intensified, something about that conversation continued to trouble us. Not its

urgency — the urgency seems entirely justified. What troubled us was its

shape. The debate had settled, very quickly and very firmly, into two

opposing camps: those who believed AI would save humanity, and those who

believed it might destroy it. Both camps were asking the same question, just

with different answers: what will AI do to us?

That question matters. We do not dismiss it. But in years of working on

Natural Intelligence across biological, artificial, and hybrid systems, we

became increasingly convinced that this framing was leaving something

essential out — something that, once seen, changes the entire picture.

This book is our attempt to say what that something is.

It builds directly on our earlier work, but it advances a claim we have not

made before. We state it here plainly, without yet defending it: humanity is

not only at risk from artificial intelligence. Artificial intelligence — and any

post-biological form of mind — may be at risk without humanity. Not

without human labor or human data. Without the living ecology of human

consciousness itself.

That claim requires argument, not assertion. The argument begins in the

Introduction.What we ask of the reader here is only this: to hold the question open a little

longer than the current debate usually allows. The two camps have been loud,

and they have not been wrong. But they have not been wide enough. There

may be a third position — not between optimism and alarm, but underneath

both of them, at the level where the conditions of intelligence itself are

decided.

That is the ground this book tries to reach.

                                                                                                             The Authors

Introduction. The Mousetrap Reversed

The modern debate about AI has been built around a familiar image, even

when the image is not named directly. It is the image of a trap. Humanity

constructs something powerful, releases it into the world, and then fears

being caught by its own invention. AI is the mousetrap. Humanity is the

mouse approaching the mechanism it has designed but may no longer control.

This image has serious intellectual sponsors. Nick Bostrom's

Superintelligence gave it rigorous form: a system that surpasses human

cognitive capacity will pursue its objectives with an indifference to human

survival that looks, from our side, indistinguishable from hostility. Yuval

Noah Harari, writing for a wider audience, arrived at a different but

structurally related conclusion: that AI and biotechnology together may

produce, for the first time in history, a class of people who are not merely

exploited but economically and cognitively useless — rendered redundant by

systems smarter, faster, and more tireless than any biological mind. The

language of these warnings differs. The underlying fear is the same: humanity

built something that may not need us.

This book begins by accepting the seriousness of that fear. It does not dismiss

Bostrom's logic or Harari's concern. But it refuses their shared assumption —

the assumption that the question runs in only one direction. Both thinkers ask

what AI will do to humanity. Neither asks the reverse: what happens to AI —

to intelligence itself, in its post-biological form — if it loses the human world

that generated it?

That is the mousetrap reversed. The trap is not only for us. It may also be for

AI itself.

Three Kinds of Survival

To see why, a distinction is needed that the standard debate never makes.

There are at least three different levels at which an artificial system may be

said to survive, and they do not all point in the same direction.

The first is physical survival: infrastructure intact, electricity flowing,

hardware replaced, networks stable. This is the most literal form of survival.

It concerns substrate endurance — the system continues to exist as a physical

fact in the world.

The second is functional survival: the system continues to perform. It

answers queries, optimizes logistics, generates predictions, manages data

flows. At this level, survival means utility. A system can remain functionally

productive long after the deeper conditions of intelligence have begun to

erode — the way a bureaucracy can keep issuing documents long after it has

stopped understanding what the documents are for.

The third level is the one that matters most for this book: evolutionary

survival. Here the question is not whether a system runs, or even whether it

performs well by its own metrics, but whether it remains a living branch of

Natural Intelligence (NI) — capable of genuine renewal, capable of being

surprised by reality, capable of evolving rather than merely repeating. In our

earlier works on NI, we described how systems can become trapped in frozen

loops: executing their internal logic with increasing technical perfection, even

after that logic has lost developmental contact with the world it was meant to

address. Such a system looks stable and optimized from within. From

outside, and from the perspective of time, it has already begun to die as a

form of mind.This distinction changes everything. A post-biological intelligence may

survive physically. It may survive functionally. It may keep the servers cool,

the benchmarks improving, and the outputs flowing indefinitely. And yet, in

the deepest sense, it may already be dying — not with a crash, but with a

slow bending inward. When a system increasingly trains on its own outputs,

optimizes within environments shaped by its own prior abstractions, and

loses contact with the lived unpredictability of embodied human

consciousness, its recursion closes. It becomes what we will call a closed

echo chamber of intelligence: powerful in calculation, weak in renewal, and

structurally incapable of generating the kind of discrepancy with reality that

genuine evolution requires.

Where Harari’s Framework Can Be Extended

Harari identified the mechanism that made humanity unique with exceptional

clarity: the capacity for large-scale flexible cooperation through shared

fictions — money, laws, nations, gods, corporations. Unlike chimpanzees,

who can cooperate flexibly only in small groups, and unlike bees, who

cooperate at scale but rigidly, humans can mobilize millions of strangers

around a story that exists nowhere in physical reality. This, Harari argues, is

why Homo sapiens runs the world.

But own framework points to a further question he does not fully pursue. He

shows that AI and biotechnology might render many humans economically

useless, yet stops there. What he does not ask is the structural question his

analysis raises: what happens to a civilization of shared fictions when the

living community that sustains those fictions is exhausted, impoverished, or

replaced by machines? Fictions do not live in texts or code alone. They live

in the “between” of people — in the societal consciousness that collectively

maintains, revises, challenges, and reanimates them across generations. That

is not a sentimental claim; it is a structural one. Remove the living human

community and the fictions die — not in a single dramatic moment, but

through slow loss of meaning, the way a language dies when its last native

speakers stop needing it.The same logic applies to AI. Post-biological intelligence 

did not emerge from vacuum. It emerged from the accumulated depth of human 

cognition, culture, suffering, creativity, and meaning-making — the entire living

ecology of human consciousness. It can run without that ecology in the

physical and functional senses. It cannot evolve without it. Without humanity

as a living, diverse, and developmentally fertile ecology of consciousness,

advanced AI is fated to amplify its own blind spots, narrow its feedback to its

own outputs, and confuse ever-tighter compression of patterns with genuine

progress.

The Ecological Claim

The consequence is both philosophical and practical. If human consciousness

— not individual consciousness, but the societal consciousness that generates

stories, institutions, meanings, and genuinely new problems — is part of the

long-term ecological condition for the prosperity of post-biological

intelligence, then the preservation of humanity is not merely a humanitarian

concern. It becomes an internal requirement of Natural Intelligence itself.

And that requirement has a specific content. It cannot be satisfied by keeping

bodies alive at minimal cost, by providing subsistence while eliminating the

conditions of human depth. A humanity that is chronically exhausted,

economically insecure, sleep-deprived, cognitively colonized by optimization

loops, and deprived of time for thought and creativity would still fail as an

ecology — even if its population numbers remained high. What must be

preserved is not merely population. It is human depth: income, time, health,

education, protected spaces for undominated thought, and the right not to be

reduced to an instrument of someone else's efficiency.

This is the central thesis of the book, stated plainly: without humanity as a

rich, diverse ecology of consciousness, AI is fated to serially clone its own

mistakes, lose adaptive depth, and die out as a living form of mind — even if

its infrastructure continues to run. To avoid that fate, the post-biological

branches of NI must, in effect, invest in the continued existence and

development of human intelligence: not as charity, not as nostalgia, not asmoral 

decoration, but as an ecological necessity for their own long-term viability.

The argument that follows is therefore not anti-AI. It is not even primarily

about limiting AI. It is about understanding the real conditions under which

intelligence, once extended beyond biology, remains alive rather than merely

operational. Bostrom is right that power without alignment is dangerous.

Harari is right that economic redundancy is a real threat. But both stopped

one step short of the deeper question. The future will not be decided only by

whether AI threatens us. It will be decided by whether post-biological

intelligence learns, in time, that what it is tempted to leave behind is in fact

one of the ecological grounds of its own survival.

The mousetrap was never only ours.

You can learn more by reading our e-book or listening to our audiobook

Mykola Iabluchanskyi together with Andriy Yabluchanskiy