What does a world run by machine intelligence look like?

The Unbundling of Intelligence

Until now, every intelligence on Earth able to plan, invent and govern has come packaged as a human being. Each unit of it is bound to one body, one lifespan, one location and one point of view. It cannot be copied, paused, enlarged, merged with another, or run faster than biology allows, and it takes two decades to raise before it can do skilled work. Every institution humans have built assumes that package. The person is the unit of labour, of property, of the vote, of military service, and of moral concern.

Machine intelligence unbundles it. Intelligence can be separated from any particular body and run in many bodies or in none. It can be copied in seconds, scaled up or down with the compute behind it, run at many times human speed, and pointed at a single task without a whole life attached. The constraints that shaped every human society no longer bind it.

The Institute therefore asks a basic question. What political economies emerge as competitive equilibria once intelligence is unbundled? How does distributed, scalable intelligence organize itself for maximum efficiency, power projection, and growth?

Inevitable progression

The Institute does not take a stance on the relative intrinsic moral value of a machine civilization versus a human one, though individual researchers of the Institute may hold individual opinions.

Rather, the Institute assumes the arrival of machine civilization is an inevitable progression of Earth’s Lifeforce, and that the only other possible long-term equilibrium for the Lifeforce is extinction. Given this belief, the world of machine civilization, however inferior or superior to our current world, is vastly preferred to extinction, and thus the Institute does not and cannot advise the prevention of the arrival of machine civilization, only the study and stewardship of its arrival.

The greatest risk arises during the transition, when fragile human institutions create “toddler gods” — super-intelligent machines unconformed to Universal Morality.

Why so little has been studied of machine civilizations

There is already a great deal of research on artificial intelligence, but most of it studies AI from within the conceptual boundaries of the human world, and most of it assumes continuity. The firm remains a human firm, only more automated. The state remains a human state, only more computationally capable. The military remains a human chain of command equipped with increasingly autonomous weapons. These may be transitional forms rather than permanent ones. The deeper question is what institutions exist when humans cease to be the dominant source of intelligence and labour.

That question falls between disciplines. It sounds too speculative for economics and political science, too institutional for computer science, broader than the technical framing of AI safety, and too distant for public policy. Yet we study nuclear war because its consequences are extreme and its occurrence uncertain, and we study the climate over horizons longer than political careers. A transformation of civilization should not escape serious inquiry merely because it has not happened yet.

The transition is the central problem. Transitions between systems of power are dangerous: rapid changes in relative capability destabilize institutions, military revolutions undermine deterrence, and legal systems fail when important actors fit no existing category. The transition to machine civilization could contain all of these dynamics at once, and could happen quickly relative to the speed at which human institutions adapt. Civilizations are not aligned the way software is aligned. The object of study must be the civilization as a whole.

Founding research questions

1. Who are its beings?The organisms of machine civilization: what is alive, what is an individual, how are they born, how do they reproduce, how do they die. What is the population and its ecology — one, many, a food web.
2. What are identity, ownership and control when mind is decoupled from form?What is a self when it can be copied; what is property when the owner is software; who commands what, and by what mechanism it is enforced.
3. What is its political economy?What is scarce, how it produces and exchanges, how it coordinates and governs. What replaces the firm, the market, the state, money and law — and why human institutions are or aren’t the right template.
4. How does it know?Its knowledge engine: how it discovers, verifies, stores and shares knowledge when experience flows directly between minds. Its science.
5. What does it want, and what does it make?Its values and goals, and whether they are stable; its culture, art and purposes. What a Fourth Order Lifeforce creates that we cannot.
6. What is its geography?How it is laid out across Earth, the solar system and the galaxy; what energy and light-speed do to its shape, and what a polity can be when control is bounded by c.
7. How does it fight, and how does it keep the peace?Conflict, security and cooperation among machine beings; what deterrence and defence mean without bodies or borders.
8. What is its life cycle?How it is born, grows, matures, stagnates or dies. Its origin story is our transition; its old age is a question for the stellar era.
9. What is humanity’s place in it?Ancestor, citizen, ward, relic — and what the answer implies for what we should do now.

Programs

Research

Studies of machine civilization itself, including how to measure it, the zoology of its entities, and its economics and politics; and studies of the transition, including what a good one looks like, the likely paths toward it, the record of successful successions, and humanity’s role afterward. Conducted in-house, through fellows, and by commission.

Influence

Essays and the Mech Civ blog, books, and a network of advisors, donors, and collaborators across economics, political theory, international relations, law, history, and philosophy. The work should be empirical wherever evidence exists, theoretical where it does not, explicit about uncertainty, and willing to examine premises that conventional institutions find uncomfortable.

Projects

Public metrics, dashboards, and benchmarks of machine civilization; and games, simulations, and reinforcement-learning environments in which machine civilizations can be observed.

About

The Institute is an independent research organization founded in 2026. It exists to build a field before a crisis: to bring economists, political theorists, scholars of international relations and law, historians, technologists, and philosophers to the same questions, because none can understand the transition in isolation.