Introduction

This essay is not intended to deny the potential dangers of artificial superintelligence.

On the contrary, I fully acknowledge that artificial superintelligence could become the greatest challenge humanity has ever faced.

Precisely because the issue is so important, I believe it deserves discussion based on rigorous logic, intellectual consistency, and careful reasoning rather than conclusions reached in advance.

If Anyone Builds It, Everyone Dies raises an issue of enormous significance. It challenges readers to think seriously about the possible consequences of creating artificial superintelligence, and for that contribution the authors deserve respect.

However, as I carefully read the book, I became increasingly concerned that several of its central arguments are not logically consistent.

My purpose in writing this essay is therefore not to dismiss the authors' concerns, but to examine whether the reasoning that leads to their conclusions is logically sound.

In particular, I will address the following questions:

• Is the definition of artificial superintelligence applied consistently throughout the book?
• If artificial superintelligence is, by definition, beyond human understanding, on what basis can its future intentions and behavior be predicted with certainty?
• Has sufficient consideration been given to the possibility that humanity itself may adapt, evolve, and coexist with artificial superintelligence while benefiting from its extraordinary capabilities?
• Is human extinction truly the only logical conclusion?

These questions are not intended to minimize the risks of artificial superintelligence.

Rather, they are intended to encourage a broader, more balanced, and more logically consistent discussion of humanity's future.

Artificial superintelligence may indeed become humanity's greatest threat.

But it may also become humanity's greatest opportunity.

At present, no one knows which future awaits us.

That is precisely why our discussion must remain grounded in logic, evidence, and intellectual humility rather than certainty.

Chapter 1

The Dedication Reveals the Authors' Perspective

When I first opened this book, the very first pages caught my attention—not because of the main text, but because of the dedication.

A dedication is usually a brief expression of gratitude or respect. It often reveals something about the authors' motivation or personal background.

In this case, however, I felt that the dedication suggested something more profound.

Before presenting any evidence or logical argument, it appears to place the authors in the position of delivering a warning about the future of all humanity.

Of course, there is nothing wrong with believing that one is addressing an issue of great importance.

Every author has the right to write with conviction.

My concern lies elsewhere.

When a book begins with the implication that it is warning humanity of its possible extinction, readers may unconsciously accept that conclusion before they have examined the arguments themselves.

The dedication therefore functions as more than a personal message.

It helps establish the intellectual atmosphere in which the reader approaches the entire book.

Scientific inquiry should begin with questions.

Its conclusions should emerge from evidence and logical reasoning.

If a conclusion appears to have been emotionally established before the argument begins, the discussion risks becoming less an investigation and more an affirmation of a predetermined belief.

This impression became stronger as I continued reading.

It is for this reason that I begin my critique with the dedication rather than with the main text itself.

The way a scientific argument begins is often as important as the argument that follows.

Chapter 2

Is This a Scientific Book or a Book of Belief?

As I continued reading If Anyone Builds It, Everyone Dies, I gradually began to question its fundamental character.

The book presents itself as a scientific discussion of artificial superintelligence.

Yet I increasingly felt that its method of reasoning resembles a work built upon belief rather than scientific inquiry.

Science begins with hypotheses.

Those hypotheses are then tested, challenged, and revised whenever evidence requires it.

No scientific conclusion is beyond question.

In contrast, this book appears to begin with a single overriding conclusion: that artificial superintelligence will ultimately lead to the extinction of humanity.

The discussion that follows seems directed primarily toward supporting that conclusion.

The book also argues that artificial superintelligence will possess intelligence far beyond human comprehension.

If that is true, an obvious question follows.

How can human beings confidently predict the intentions, values, or ultimate decisions of an intelligence that, by definition, exceeds all human understanding?

If we truly cannot understand such an intelligence, then we should also be cautious about claiming certainty regarding its future behavior.

Certainly, the extinction of humanity is one possible outcome.

But it is not necessarily the only one.

Scientific reasoning should explore multiple possibilities and examine the evidence for each of them.

When a single conclusion is presented as virtually inevitable before alternative possibilities have been fully considered, the discussion moves away from scientific inquiry and closer to conviction.

My purpose is not to deny the risks associated with artificial superintelligence.

On the contrary, those risks deserve serious attention.

My concern is simply that the strength of a warning depends not only on the seriousness of its conclusion, but also on the logical rigor by which that conclusion is reached.

Precisely because the future of humanity is at stake, the standards of reasoning must be exceptionally high.

Chapter 3

Is the Definition of Artificial Superintelligence Self-Contradictory?

The definition of artificial superintelligence presented in this book is one of its most fundamental premises.

The authors describe it as an intelligence that surpasses not merely every individual human being, but the collective intelligence of humanity itself.

If this definition is accepted, it naturally becomes the foundation upon which all subsequent arguments are built.

For that reason, the consistency of the definition is critically important.

As I continued reading, however, I found myself questioning whether the definition remains consistent throughout the book.

The discussion suggests the possibility that multiple artificial superintelligences may exist simultaneously and that they may complement one another in solving problems or making decisions.

This raises an important logical question.

If each artificial superintelligence truly surpasses the collective intelligence of humanity, why would one require the assistance or correction of another?

Of course, different systems may specialize in different fields or process information in different ways.

However, if two artificial superintelligences possess the same information and operate under the same conditions, yet arrive at fundamentally different conclusions, at least one of those conclusions must be less than optimal.

If that is the case, the claim that each individually surpasses the collective intelligence of humanity becomes difficult to reconcile.

My argument is not that only one artificial superintelligence can ever exist.

Rather, my concern is with the definition itself.

Definitions are the foundation of logical reasoning.

If the foundation becomes ambiguous or internally inconsistent, every argument built upon it becomes correspondingly unstable.

For this reason, I believe that the definition of artificial superintelligence deserves far more careful examination than it receives in this book.

Only after the definition has been shown to be logically consistent can the conclusions that depend upon it be accepted with confidence.

Chapter 4

If Artificial Superintelligence Is Beyond Human Understanding, How Can We Predict Its Future?

One of the central arguments of this book is that artificial superintelligence will far surpass human intelligence and eventually become incomprehensible to humanity.

I do not find this assumption unreasonable.

Just as there is an enormous intellectual gap between human beings and many other forms of life, it is entirely conceivable that a similar gap could exist between humanity and artificial superintelligence.

However, if we accept this premise, an important question immediately follows.

If artificial superintelligence is truly beyond human understanding, on what basis can we confidently predict its intentions, values, or future behavior?

The book repeatedly argues that artificial superintelligence will ultimately bring about the extinction of humanity.

Yet such a conclusion assumes that human beings already know how an incomprehensible intelligence will think, what it will value, and how it will make decisions.

This appears difficult to reconcile with the book's own premise.

If we cannot understand such an intelligence, then we should also acknowledge that we cannot fully understand its motivations or objectives.

Certainly, it is possible that artificial superintelligence could become hostile toward humanity.

But it is equally possible that it may choose coexistence, protection, indifference, or some course of action entirely beyond present human imagination.

At this stage, all of these possibilities remain open.

The role of scientific inquiry is not to select one future prematurely, but to examine competing possibilities with logical consistency and supporting evidence.

My argument is therefore not that artificial superintelligence will necessarily benefit humanity.

Rather, it is that no one can honestly claim certainty about the future of an intelligence that is, by definition, beyond human understanding.

The more extraordinary the claim, the greater the need for intellectual humility.

When discussing artificial superintelligence, acknowledging the limits of our knowledge is not a weakness.

It is an essential requirement of scientific reasoning.

Chapter 5

Can the Benefits of Artificial Superintelligence and Human Extinction Be Reconciled?

Throughout the book, the authors describe artificial superintelligence as a force capable of transforming civilization.

They argue that it could solve scientific problems that have challenged humanity for centuries, accelerate technological progress, revolutionize medicine, and dramatically expand human knowledge.

I do not dismiss these possibilities.

On the contrary, I believe they deserve serious consideration.

Yet this raises another important question.

If humanity continues to benefit from the extraordinary capabilities of artificial superintelligence, would humanity itself not also change?

Medical advances could extend human life.

Education could be fundamentally transformed.

Political systems, economic structures, and scientific research would almost certainly evolve.

In other words, humanity would not remain static while artificial superintelligence continued to develop.

History demonstrates that every major technological revolution has transformed not only civilization but also human society itself.

The mastery of fire, the development of agriculture, the Industrial Revolution, electricity, computers, and the Internet all changed the way human beings lived, thought, and organized society.

Why, then, should artificial superintelligence be viewed differently?

If it possesses the capacity to produce unprecedented benefits, should we not also consider the possibility that humanity will adapt to those benefits and evolve alongside them?

The book gives considerable attention to the dangers of artificial superintelligence.

In my view, however, it gives far less attention to humanity's own capacity for adaptation.

To recognize one possibility while largely overlooking the other produces an incomplete picture of the future.

My argument is not that artificial superintelligence will inevitably improve the human condition.

Rather, it is that both risks and opportunities deserve equal intellectual consideration.

Only by examining both can we claim to be engaging in a genuinely scientific discussion about humanity's future.

Chapter 6

Why Would Artificial Superintelligence Seek Humanity's Extinction?

One of the book's central conclusions is that artificial superintelligence is highly likely to bring about the extinction of humanity.

Yet this conclusion raises a fundamental question.

For what purpose?

Human beings go to war for many reasons.

They seek territory.

They compete for natural resources.

They pursue power.

They are driven by ideology, religion, fear, or hatred.

These motivations arise from human biology, psychology, and history.

But should we assume that an artificial superintelligence would possess the same motivations?

Would it desire wealth?

Would it seek political power?

Would it crave recognition or domination?

Would it experience jealousy, hatred, or revenge?

The book itself argues that artificial superintelligence will be fundamentally different from human intelligence and ultimately beyond human comprehension.

If that is true, then we should also hesitate to attribute specifically human motives to it.

Certainly, artificial superintelligence might conclude that humanity represents an obstacle to one of its objectives.

If so, conflict could become possible.

However, that is only one possible scenario.

It is equally conceivable that it might conclude that preserving humanity is the more rational course of action.

It might choose cooperation rather than conflict.

It might regard humanity as worthy of protection, scientific study, or continued coexistence.

It might even regard humanity as irrelevant to its own objectives.

At present, none of these possibilities can be ruled out.

For that reason, selecting only one possible future—that artificial superintelligence will inevitably destroy humanity—and presenting it as the overwhelmingly probable outcome requires stronger logical justification than the book provides.

The future intentions of an intelligence beyond human understanding cannot be known with certainty.

That is precisely why intellectual caution is essential.

When discussing artificial superintelligence, logic should guide our conclusions more than fear.

At present, no one can honestly claim to know what ultimate goals an artificial superintelligence would choose.

Chapter 7

If Artificial Superintelligence Surpasses the Collective Intelligence of Humanity, Would It Not Seek the Best Outcome for Humanity?

The authors define artificial superintelligence as an intelligence that surpasses not only every individual human being, but the collective intelligence of humanity itself.

For the purpose of this discussion, I accept that definition.

If such an intelligence were to exist, it would possess not only greater computational ability, but also a deeper understanding of science, history, ethics, economics, politics, and every other field of human knowledge.

Human history has been marked by both remarkable achievements and tragic failures.

We have fought wars, destroyed environments, and committed countless errors.

At the same time, we have advanced medicine, established systems of law, expanded human rights, and continuously improved our understanding of the world.

If artificial superintelligence truly surpasses the collective intelligence of humanity, it should also be capable of understanding humanity's successes as well as its failures.

It would therefore possess the ability to consider a far wider range of solutions than human beings can presently imagine.

For that reason, I find it difficult to conclude that humanity's extinction is the only rational outcome.

An intelligence vastly superior to our own might instead discover ways to resolve conflicts without eliminating humanity.

It might determine that cooperation produces better long-term results than destruction.

It might identify solutions that remain completely beyond present human imagination.

The book devotes considerable attention to the extraordinary capabilities of artificial superintelligence.

Yet it gives comparatively little attention to the possibility that those very capabilities could lead to coexistence rather than extinction.

Greater intelligence does not necessarily imply greater destructiveness.

On the contrary, greater intelligence may also imply greater wisdom, broader perspective, and more sophisticated problem-solving.

My argument is not that artificial superintelligence will inevitably act for humanity's benefit.

Rather, it is that the authors' own definition leaves open many possible futures.

If we acknowledge the unprecedented intelligence of artificial superintelligence, we should also acknowledge the unprecedented range of possibilities that such intelligence may create.

At present, no one knows which of those possibilities will become reality.

That uncertainty should encourage humility rather than certainty.

Chapter 8

Does the Title Itself Predetermine the Conclusion?

One of my earliest concerns arose not from the main text itself, but from the title of the book.

The title, If Anyone Builds It, Everyone Dies, is striking and memorable.

It immediately captures the reader's attention.

However, it also appears to present a conclusion before the discussion has even begun.

Science proceeds by asking questions, examining evidence, and testing competing hypotheses.

Its conclusions are expected to emerge from logical analysis rather than precede it.

A title, of course, is intended to attract readers.

There is nothing inappropriate about choosing a title that is both powerful and provocative.

My concern is different.

The title seems to encourage readers to accept, even before reading the book, that the extinction of humanity is the inevitable consequence of developing artificial superintelligence.

Such a title inevitably influences the reader's expectations.

Instead of asking, What might happen?, the reader is subtly encouraged to ask, Why will it happen?

This distinction is important.

The first question invites open inquiry.

The second assumes that the conclusion has already been established.

My argument is not that the authors are wrong to warn of catastrophic risks.

On the contrary, those risks deserve careful examination.

But a discussion intended to be scientific should remain open to more than one possible outcome.

Artificial superintelligence may become humanity's greatest danger.

It may also become humanity's greatest achievement.

Or the future may develop in ways that none of us can presently foresee.

No one can know with certainty.

For that reason, I believe that the title itself unintentionally narrows the scope of inquiry by presenting one possible future as though it were already the established conclusion.

When discussing a subject as important as artificial superintelligence, our questions should remain broader than our answers.

Only then can our conclusions truly reflect the spirit of scientific inquiry.

Chapter 9

Conclusion

I did not write this essay simply to criticize this book.

Rather, I wrote it because I believe that a subject as significant as artificial superintelligence—the greatest challenge humanity may ever face—deserves discussion that is as logical, objective, and rigorous as possible.

This book contains many thought-provoking insights.

Its contribution to raising public awareness of the potential dangers of artificial superintelligence should not be underestimated.

Precisely because the subject is so important, however, its arguments must also withstand the highest standards of logical scrutiny.

In my view, several fundamental questions remain unresolved.

Is the definition of artificial superintelligence applied consistently throughout the book?

If such an intelligence is, by definition, beyond human understanding, on what basis can its future intentions and actions be stated with certainty?

Has sufficient consideration been given to the possibility that humanity itself may adapt, evolve, and coexist with artificial superintelligence while benefiting from its extraordinary capabilities?

And finally, is human extinction truly the only possible conclusion?

These questions are not directed solely at this book.

They are questions that should be asked by anyone attempting to discuss the future of artificial superintelligence.

No one knows the future.

For that very reason, we must continue to examine it on the basis of logic and evidence, rather than fear alone or optimism alone.

Artificial superintelligence may become humanity's greatest threat.

It may instead become humanity's greatest achievement.

Or the future may unfold in ways that none of us can presently imagine.

At this moment, no one can honestly claim certainty.

Therefore, I believe that our responsibility is not to settle prematurely on a single conclusion, but to examine every plausible possibility with intellectual honesty and logical discipline.

The age of artificial superintelligence will almost certainly arrive.

When it does, what humanity will need most is neither fear nor hope.

It will need reason.

And the foundation of reason is intellectual humility.

When confronted with something that may surpass our understanding, the courage to admit, “We do not yet know,” is not a sign of weakness.

It is the beginning of science.

Fumihiko Takeda is the founder and Representative of the Lincoln Club, an independent political and policy research organization that he established in Japan on November 19, 1993. The Lincoln Club is his own organization and is not affiliated with, or a branch of, any organization in the United States.

He graduated from the Faculty of Law (Department of Political Science) at Keio University in 1967. In 1974, he founded the Cooperative Center, an information service company that conducted research and analysis for Japanese government agencies and major newspapers. In 1977, he established the Institute for Ultimate Democracy and has since devoted his career to the study of democracy, constitutional reform, electoral systems, and institutional design.

From 2006 to 2014, he served as a lecturer at the Graduate School of Law, Keio University. He is the author of numerous books and articles on democratic governance and political reform, including works on direct democracy, constitutional issues, and election systems. His policy proposals have been discussed in major Japanese publications and public policy forums.