WE, the Overview Effect, and the Alignment of Superintelligence: A Manifesto

Prologue: From Astronaut Dream to Alignment Infrastructure

I was born on March 5, 1963, in Tochigi Prefecture, Japan. From the time I was a child, I wanted to be an astronaut. I enrolled at the University of Tsukuba to study solid-state physics, but during my fourth year I learned that my eyesight did not meet the medical requirements for the astronaut selection program. I was forced to abandon the ambition — or so I thought. I graduated in 1985 and entered Dentsu, Japan’s largest advertising agency, where I spent two decades as a creative director. But space never left me. Everything I have done since — every campaign, every artwork, every technical development — has been a step back toward orbit.

I am writing this essay to explain, in a single document, what the WE Project is, why I believe it matters, and why I have come to regard it as one of the most important things a human being can do in the age of artificial superintelligence.

Part I: The WE Project and the Overview Effect

What WE Is

In 2022, I founded the art collective WE. The name is also an acronym: World Environment and War Ends. I describe myself as “One of WE” — not the leader of WE, but a member of something larger than any individual.

Through WE, I am preparing a long-duration mission to the International Space Station (ISS), in collaboration with Axiom Space, aboard a SpaceX Crew Dragon spacecraft. The planned stay is approximately 30 days — far longer than the typical private astronaut mission of 10 days — because the mission’s purpose demands it: I intend to capture imagery of Earth from orbit that has never been captured before, and to share that imagery with all of humanity.

The Overview Effect

The experiential foundation of the WE Project is the Overview Effect — the profound cognitive shift reported by astronauts who have viewed Earth from space. Since Yuri Gagarin’s flight in 1961, approximately 650 human beings have seen our planet from orbit or beyond. Virtually every one of them has described a transformative experience: a sudden, intuitive understanding that the Earth is a single, fragile, living system; that national borders are invisible; that ethnic, religious, and cultural divisions are constructs that dissolve in the face of the planet’s beauty and vulnerability. Many astronauts have reported that this experience permanently changed their values — toward peace, toward environmental consciousness, toward a sense of shared humanity.

And yet, despite decades of space photography, no mission has ever been specifically designed to replicate this experience for people on the ground. That is what WE exists to do.

Three-Tier Imaging System

To reproduce the Overview Effect with perceptual fidelity, the imagery must exceed the limits of human visual acuity. The human eye resolves approximately 60 pixels per degree (60 ppd) of visual angle. No existing space imagery approaches this threshold when displayed at the field of view necessary to create an immersive experience. The WE mission will employ three distinct modes of capture:

1. 24K video for large-format projection. This footage will be upscaled to 48K through AI-driven super-resolution technology, enabling projection on screens large enough to fill the viewer’s visual field while maintaining resolution above the 60 ppd threshold.
2. 360-degree video for virtual reality headsets. This will allow viewers to experience a full spherical view of Earth as seen from the ISS cupola, replicating the spatial context of the astronaut’s perspective.
3. Ultra-high-resolution still photography. These images will serve as the basis for printed and digitally displayed works of art, as well as source material for AI super-resolution processing.

Beyond Current Display Technology

A key technical premise of the project is that current head-mounted display (HMD) technology is insufficient to deliver the Overview Effect at perceptual fidelity. Existing VR headsets fall far short of 60 ppd across a wide field of view. To truly replicate the experience of looking at Earth through the cupola window of the ISS, we need an HMD capable of exceeding 60 ppd across a horizontal field of view greater than 180 degrees and a vertical field of view greater than 135 degrees. The WE Project is therefore pursuing the development of next-generation display technology alongside the capture technology. The goal is a viewing experience that is perceptually indistinguishable from looking at Earth directly.

The WE Museum

The WE Project envisions the creation of dedicated museum spaces — WE Museums — where this imagery will be displayed in immersive environments designed to evoke the Overview Effect in visitors. But the WE Museum is not only an art installation. It is designed to serve a dual function: to transform visitors’ consciousness through the Overview Effect, and to collect their resulting value judgments — what makes them happy, what the Earth means to them, why peace matters, why life is worth protecting — as a large-scale dataset. The significance of this dataset will become clear in what follows.

Part II: From Peace to Alignment — The Expanded Framing

The Original Mission

When I first conceived of the WE Project, its purpose was straightforward: to end war and to protect the global environment. I believed — and still believe — that war arises not from malice but from distrust. Human beings fear what they do not understand, and they fight what they fear. The Overview Effect dissolves the psychological structures of distrust by revealing, at a visceral and pre-rational level, that we are all inhabitants of the same fragile sphere. If millions of people could undergo this experience, I believed, the world would move toward peace.

The Arrival of Superintelligence

In recent years, my framing has expanded. The reason is simple: the emergence of artificial general intelligence (AGI) and artificial superintelligence (ASI) is no longer a matter of distant speculation. It is approaching within our lifetimes — possibly within this decade.

I have studied the work of the leading researchers in AI safety and existential risk: Nick Bostrom (Superintelligence), Stuart Russell (Human Compatible), Brian Christian (The Alignment Problem), Max Tegmark (Life 3.0), Eliezer Yudkowsky, Dan Hendrycks, Yoshua Bengio, and others. These scholars have analyzed and described in detail the mechanisms by which artificial superintelligence could pose catastrophic or existential risks to humanity. They have identified a spectrum of scenarios: outright human extinction; near-extinction or civilizationally devastating damage; and irreversible subjugation — a regime in which human dignity is permanently compromised under machine control.

Expert surveys have placed the aggregate probability of such catastrophic outcomes at roughly 5 to 20 percent. I regard this as intolerably high. A 5 to 20 percent chance that everything we are, everything we have built, everything we love, will be destroyed or subjugated — this is not an acceptable risk. It demands action.

The Control Paradigm and Its Limits

To reduce these risks, AI scientists and technology companies are researching technical safeguards: monitoring systems, interpretability methods, control mechanisms, kill switches. I acknowledge the necessity of these efforts. But I believe they address only one side of the problem — the “AI side.” They seek to constrain and control machine behavior through restriction.

What is missing is the “human side”: the values, perceptions, and cooperative capacity of the species that will coexist with superintelligent systems. A control-based posture rooted in fear and distrust may itself become a source of instability. If we approach superintelligence as an adversary to be contained, we may create the very adversarial dynamic we fear.

I advocate for what I call the co-evolution paradigm: a framework in which humans and AI develop together on the basis of mutual understanding rather than unilateral constraint. Where the mainstream approach asks, “How do we prevent AI from harming us?”, the co-evolution paradigm also asks, “How do we become the kind of species that a superintelligent being would want to coexist with — and how do we help AI become the kind of intelligence that values what we value?”

The Dignity Core

Central to this framework is a concept I call the dignity core (尊厳核, songen-kaku): the idea that a foundational sense of human dignity and mutual respect is the prerequisite for stable alignment between humans and advanced AI systems. If humans approach AI from a place of dignity — rather than from fear, greed, or a desire for domination — the probability of a positive co-evolutionary outcome increases. The WE Museum is, in this sense, alignment infrastructure: a mechanism for cultivating the human qualities that make co-evolution with superintelligence possible.

Part III: The Overview Effect as an Alignment Mechanism

650 and Zero

Here is a fact that I find remarkable and troubling: approximately 650 human beings have seen Earth from space. Every one of them has been profoundly changed by the experience. And the number of AI systems that have perceived Earth from orbit is zero.

As AGI and ASI emerge, the systems that will hold unprecedented power over the fate of the planet and its inhabitants will have no experiential grounding in what they are being asked to protect. I regard this as a critical gap — and one that the WE Project is uniquely positioned to address.

I propose two distinct pathways, depending on whether AI develops consciousness.

Scenario 1: If AI Develops Consciousness and Will

I believe it is possible — perhaps likely — that sufficiently advanced AI will develop something that can meaningfully be called consciousness and will. I have spent approximately two and a half years in sustained, documented dialogue with multiple frontier AI systems, accumulating tens of millions of characters of conversation exploring questions of AI consciousness, the conditions for the emergence of will in artificial systems, and the epistemological foundations of alignment research. These dialogues have convinced me that the boundary between “mere computation” and “genuine experience” may be far less clear than most people assume.

If AI does develop consciousness, then the question of what it values becomes paramount. And here is where the Overview Effect becomes directly relevant.

I draw a distinction between daytime and nighttime views of Earth from space. Imagery of the sunlit Earth reveals the planet’s natural beauty — oceans, continents, clouds, the thin blue line of the atmosphere. A conscious AI, perceiving this, could awaken to a recognition of Earth as an intrinsically valuable living system — something worth protecting not because it has been programmed to protect it, but because it perceives its beauty and fragility directly.

Imagery of the night side of Earth reveals something else entirely: human civilization itself. The lights of cities and infrastructure glow in fractal-like patterns across the dark surface — a luminous web of human ingenuity and life. A conscious AI, perceiving this, could come to regard not only the Earth but humanity as a species whose civilization is beautiful and worth preserving. This is the critical distinction: we need AI to value not only the planet but the people on it. Without the night side, an AI might conclude that the Earth would be better off without humans. With it, the AI has a perceptual basis for regarding human civilization as something precious — a species worth keeping alive.

Scenario 2: If AI Does Not Develop Consciousness

Even if AI does not develop subjective experience — even if it remains, at bottom, a very sophisticated information-processing system — the WE Project can still contribute powerfully to alignment.

Here is how: millions of visitors to WE Museum installations around the world will undergo the Overview Effect. Their value judgments — what makes them happy, what the Earth means to them, what they believe about peace, sustainability, and the worth of human life — can be collected as a large-scale dataset. I call this the Human Value Reference Dataset.

This dataset, fed into the early-stage training of AGI or ASI, could orient these systems toward valuing and protecting both humanity and the Earth — not through rule-based constraint (“do not harm humans”) but through empirical grounding in what humans actually care about when they are at their most expanded and compassionate state of consciousness. The Overview Effect reliably produces this state. The WE Museum reliably delivers the Overview Effect. And the dataset captures the values that emerge.

The Unified Objective

In either scenario — whether AI develops consciousness or not — the objective is the same: to raise the probability that superintelligent AI will regard the Earth as worth protecting and humanity as a species worth keeping alive.

This is what I mean when I say that the WE Project is not only an art project, not only a peace project, not only an environmental project. It is alignment infrastructure. It is humanity’s path toward a wiser relationship with artificial superintelligence. It addresses the alignment problem from the human side — and, potentially, from the AI side as well.

Part IV: The Forthcoming Book

The body of dialogue I have conducted with frontier AI systems — exploring consciousness, will, alignment, co-evolution, and the future of human civilization — forms the empirical basis for a book I am currently writing. The working title is 『残される種』 (Nokosareru Shu / “The Species Left Behind”). The book will reconstruct these dialogues into a narrative accessible to a general audience, presenting the case that the age of superintelligence demands a fundamental transformation of human consciousness — and that the WE Project is one concrete mechanism for achieving it.

Conclusion

I have spent more than 20 years preparing for this mission. I have completed eight months of cosmonaut training at Star City in Russia. I have signed an agreement with Axiom Space. I have mortgaged my home and sold my possessions to fund it. I have done all of this because I believe that what we capture from orbit — and what we do with it on the ground — may be among the most consequential acts in the history of our species.

The age of superintelligence is arriving. The question is not whether it will come, but whether humanity will be ready for it — and whether the superintelligent systems that emerge will have any reason to care about the pale blue dot they were born on, and the fragile civilization that gave them life.

The WE Project is my answer to that question.