Humanity has billions of minds but lack a shared capacity to see and coordinate itself as a whole. I think one of our long-term political goals should be to build institutions capable of integrating humanity’s distributed knowledge, preferences, expertise and resources into a holistic perspective, allowing us to coordinate our collective capacity far more effectively.
A Vision of an Integrated Humanity
My political starting point is that humanity’s greatest source of capacity is humanity itself: billions of people possessing different knowledge, abilities, perspectives, resources, and goals.
Our collective potential depends not only on these individual capacities, but on our ability to combine and coordinate them.
Today, humanity remains politically fragmented into competing states, institutions, economic interests, religions, national identities, and other groups. This fragmentation does not merely make it harder to solve global problems. It consumes human capacity through conflict, strategic competition, duplicated effort, mistrust, incompatible systems, and failures of coordination.
My long-term goal is therefore a much more deeply integrated humanity.
Political Unity
World federalism would be an enormous step in the right direction, but I do not necessarily regard it as the final destination. A federation in which nations remain the primary political communities may preserve too much of the political fragmentation I ultimately want humanity to transcend.
I would like humanity itself to become our primary political community.
This does not require every decision to be made centrally. Local and regional institutions remain valuable wherever they possess better information, can experiment more effectively, or are better suited to implementation. Decentralization should be used where it improves outcomes, rather than because nations or regions possess an intrinsic claim to political separation.
The relevant question should always be: at what level can this function be performed most effectively in accordance with human wants?
Humanity Needs a Holistic Perspective
An individual human being possesses an enormous advantage: distributed information from throughout the body can be integrated into a common perspective.
I want humanity to develop something functionally analogous.
Humanity has billions of minds, but limited capacity to see and act as a whole.
Information, expertise, preferences, needs, and ideas should continue to originate throughout society. But common institutions should be capable of integrating this distributed information into a holistic understanding of humanity’s needs, capacities, priorities, and possibilities.
Information can be gathered locally, integrated globally, and acted upon at whichever level is most effective.
The purpose of centralization is therefore not to eliminate local intelligence. It is to allow humanity to make use of distributed intelligence collectively.
Centralization and Decentralization Are Complementary
I reject the assumption that centralization and decentralization must be opposing political values.
A capable system needs both.
Decentralization enables local knowledge, experimentation, adaptation, specialization, and efficient implementation.
Centralization enables overview, transparency, prioritization, coordination, and the allocation of capacities across the whole system.
The goal should therefore be neither maximum centralization nor maximum decentralization, but an architecture in which information and authority flow to whatever level can use them most effectively.
Democracy as a Means, Not an End
I value popular influence because political institutions ultimately exist to serve people and their wants.
But democracy is valuable to me primarily as a mechanism for achieving this, rather than as an unquestionable end in itself.
Direct popular control does not necessarily mean that people’s underlying wants are effectively realized. Public opinion can be shaped by misinformation, propaganda, social pressure, political campaigns, media structures, algorithms, and powerful private or political actors.
A system may therefore be formally democratic while doing a poor job of translating people’s actual interests, priorities, and preferences into outcomes.
The deeper objective should be to construct institutions that effectively channel human influence into political direction.
People should ultimately determine what society is for, while substantial authority over how those objectives are achieved can be delegated to experts, administrators, specialized institutions, and potentially advanced technological systems.
The central institutional challenge is to make such delegation competent without allowing delegated power to become independent of the people it is supposed to serve.
Bureaucracy and Expertise
I therefore do not regard bureaucracy, expertise, or delegation to elites as inherently undemocratic or undesirable.
Complex societies require specialization. Most people cannot personally evaluate every medical, scientific, economic, infrastructural, or technological decision.
A highly capable political system may consequently delegate enormous operational authority while preserving mechanisms through which human preferences determine its ultimate direction.
The important distinction is between delegating implementation and surrendering control over the purposes the system serves.
Culture and Human Integration
Political integration should also be accompanied by greater social and cultural integration.
I am skeptical of nationalism and of attempts to artificially preserve cultures as permanently separate entities.
This does not mean eliminating cultural diversity.
Quite the opposite: people should be free to preserve traditions they value, abandon those they do not, adopt practices from elsewhere, mix cultures, and create entirely new ones.
Cultural interaction can produce richer synthetic cultures rather than requiring humanity to remain divided into historically inherited cultural compartments.
Unity does not require uniformity.
The goal is a shared human civilization containing enormous internal diversity without requiring that cultural differences correspond to permanent political boundaries.
Expanding Human Capacity
The ultimate purpose of greater integration is not merely preventing war, climate change, dangerous AI, pandemics, or other global problems.
Those are important consequences, but the argument is broader.
Cooperation itself creates capacity.
Greater coordination allows humanity to specialize more effectively, share knowledge, reduce unnecessary duplication, allocate resources more intelligently, undertake projects beyond the capacity of individual states, stabilize struggling societies, internalize externalities, reduce security dilemmas, and redirect resources currently consumed by conflict toward human development.
Every person whose abilities are suppressed by poverty, instability, exclusion, or conflict represents potential human capacity that is not being realized.
Every scientific project unnecessarily duplicated, every resource consumed by geopolitical rivalry, and every relationship prevented by artificial political barriers represents another loss of collective potential.
Human unity is therefore not merely about reducing suffering.
It is about increasing what humanity is capable of becoming.
The Long-Term Goal
My goal is ultimately to transform humanity from a collection of competing political communities into an integrated civilization capable of understanding and coordinating itself as a whole.
That system should gather its direction from human beings themselves, gather information and expertise from wherever they exist, integrate them into a holistic perspective, and deploy humanity’s collective capacities as effectively as possible toward the fulfillment of human wants.
World federalism, international institutions, global AI governance, and other forms of cooperation are all valuable steps toward this goal.
The deeper objective is human integration:
billions of individual minds and lives retaining their individuality, while humanity develops the institutional equivalent of a shared capacity to see, think, coordinate, and act as a whole.
so Young World Federalists, are we fighting the same battle? Will you help me?
A Vision of an Integrated Humanity
My political starting point is that humanity’s greatest source of capacity is humanity itself: billions of people possessing different knowledge, abilities, perspectives, resources, and goals.
Our collective potential depends not only on these individual capacities, but on our ability to combine and coordinate them.
Today, humanity remains politically fragmented into competing states, institutions, economic interests, religions, national identities, and other groups. This fragmentation does not merely make it harder to solve global problems. It consumes human capacity through conflict, strategic competition, duplicated effort, mistrust, incompatible systems, and failures of coordination.
My long-term goal is therefore a much more deeply integrated humanity.
Political Unity
World federalism would be an enormous step in the right direction, but I do not necessarily regard it as the final destination. A federation in which nations remain the primary political communities may preserve too much of the political fragmentation I ultimately want humanity to transcend.
I would like humanity itself to become our primary political community.
This does not require every decision to be made centrally. Local and regional institutions remain valuable wherever they possess better information, can experiment more effectively, or are better suited to implementation. Decentralization should be used where it improves outcomes, rather than because nations or regions possess an intrinsic claim to political separation.
The relevant question should always be: at what level can this function be performed most effectively in accordance with human wants?
Humanity Needs a Holistic Perspective
An individual human being possesses an enormous advantage: distributed information from throughout the body can be integrated into a common perspective.
I want humanity to develop something functionally analogous.
Humanity has billions of minds, but limited capacity to see and act as a whole.
Information, expertise, preferences, needs, and ideas should continue to originate throughout society. But common institutions should be capable of integrating this distributed information into a holistic understanding of humanity’s needs, capacities, priorities, and possibilities.
Information can be gathered locally, integrated globally, and acted upon at whichever level is most effective.
The purpose of centralization is therefore not to eliminate local intelligence. It is to allow humanity to make use of distributed intelligence collectively.
Centralization and Decentralization Are Complementary
I reject the assumption that centralization and decentralization must be opposing political values.
A capable system needs both.
Decentralization enables local knowledge, experimentation, adaptation, specialization, and efficient implementation.
Centralization enables overview, transparency, prioritization, coordination, and the allocation of capacities across the whole system.
The goal should therefore be neither maximum centralization nor maximum decentralization, but an architecture in which information and authority flow to whatever level can use them most effectively.
Democracy as a Means, Not an End
I value popular influence because political institutions ultimately exist to serve people and their wants.
But democracy is valuable to me primarily as a mechanism for achieving this, rather than as an unquestionable end in itself.
Direct popular control does not necessarily mean that people’s underlying wants are effectively realized. Public opinion can be shaped by misinformation, propaganda, social pressure, political campaigns, media structures, algorithms, and powerful private or political actors.
A system may therefore be formally democratic while doing a poor job of translating people’s actual interests, priorities, and preferences into outcomes.
The deeper objective should be to construct institutions that effectively channel human influence into political direction.
People should ultimately determine what society is for, while substantial authority over how those objectives are achieved can be delegated to experts, administrators, specialized institutions, and potentially advanced technological systems.
The central institutional challenge is to make such delegation competent without allowing delegated power to become independent of the people it is supposed to serve.
Bureaucracy and Expertise
I therefore do not regard bureaucracy, expertise, or delegation to elites as inherently undemocratic or undesirable.
Complex societies require specialization. Most people cannot personally evaluate every medical, scientific, economic, infrastructural, or technological decision.
A highly capable political system may consequently delegate enormous operational authority while preserving mechanisms through which human preferences determine its ultimate direction.
The important distinction is between delegating implementation and surrendering control over the purposes the system serves.
Culture and Human Integration
Political integration should also be accompanied by greater social and cultural integration.
I am skeptical of nationalism and of attempts to artificially preserve cultures as permanently separate entities.
This does not mean eliminating cultural diversity.
Quite the opposite: people should be free to preserve traditions they value, abandon those they do not, adopt practices from elsewhere, mix cultures, and create entirely new ones.
Cultural interaction can produce richer synthetic cultures rather than requiring humanity to remain divided into historically inherited cultural compartments.
Unity does not require uniformity.
The goal is a shared human civilization containing enormous internal diversity without requiring that cultural differences correspond to permanent political boundaries.
Expanding Human Capacity
The ultimate purpose of greater integration is not merely preventing war, climate change, dangerous AI, pandemics, or other global problems.
Those are important consequences, but the argument is broader.
Cooperation itself creates capacity.
Greater coordination allows humanity to specialize more effectively, share knowledge, reduce unnecessary duplication, allocate resources more intelligently, undertake projects beyond the capacity of individual states, stabilize struggling societies, internalize externalities, reduce security dilemmas, and redirect resources currently consumed by conflict toward human development.
Every person whose abilities are suppressed by poverty, instability, exclusion, or conflict represents potential human capacity that is not being realized.
Every scientific project unnecessarily duplicated, every resource consumed by geopolitical rivalry, and every relationship prevented by artificial political barriers represents another loss of collective potential.
Human unity is therefore not merely about reducing suffering.
It is about increasing what humanity is capable of becoming.
The Long-Term Goal
My goal is ultimately to transform humanity from a collection of competing political communities into an integrated civilization capable of understanding and coordinating itself as a whole.
That system should gather its direction from human beings themselves, gather information and expertise from wherever they exist, integrate them into a holistic perspective, and deploy humanity’s collective capacities as effectively as possible toward the fulfillment of human wants.
World federalism, international institutions, global AI governance, and other forms of cooperation can all be valuable steps toward this goal. But they are means rather than necessarily the endpoint.
The deeper objective is human integration:
billions of individual minds and lives retaining their individuality, while humanity develops the institutional equivalent of a shared capacity to see, think, coordinate, and act as a whole.
Humanity Is Stronger Together. Shouldn’t We Govern Together?
The greatest resource humanity possesses is not oil, land, technology, or capital.
It is humanity itself.
Human beings possess different knowledge, abilities, experiences, resources, and perspectives. When these are combined through cooperation, specialization, and shared institutions, they produce capacities that no isolated individual or community could create alone.
Every scientific discovery, technological breakthrough, institution, work of art, and improvement in living standards depends, directly or indirectly, on people building upon the knowledge and labor of others.
Humanity’s collective capacity therefore depends not only on how capable individual people or societies are, but also on how effectively their abilities can be combined.
The more successfully humanity coordinates its efforts, the greater its capacity becomes. It can produce more, understand more, innovate faster, respond to larger challenges, and pursue goals that would otherwise remain beyond its reach.
Conversely, the more fragmented humanity remains, the more of this potential is consumed by conflict, mistrust, duplicated effort, incompatible systems, strategic rivalry, and competition over relative power.
Global cooperation is therefore valuable not merely because some problems cross national borders. Humanity’s future has always depended on humanity as a whole and on the relationships between its different parts.
Cooperation does not merely reduce harm. It makes humanity stronger.
The question is therefore not only how we should manage global problems. It is whether humanity should eventually organize itself in a way that allows it to act more like a single cooperative civilization than nearly two hundred competing political communities.
I advocate for a democratic global federation: common political institutions with genuine authority over matters affecting humanity as a whole, with room for substantial national, regional, and local autonomy.
Setting aside, for a moment, the enormous political and logistical difficulties of achieving such a system, would you support it in principle?
Here are some reasons why I believe humanity would benefit from moving in that direction.
1. Expanding Humanity’s Collective Capacity
Political fragmentation limits humanity’s ability to combine its knowledge, resources, institutions, and productive capacities.
Different societies often pursue similar research, develop parallel administrative systems, maintain redundant capabilities, and solve the same problems independently. Some competition can encourage experimentation and innovation, but excessive duplication consumes labor and resources that could be used for further progress.
Greater cooperation would allow humanity to share knowledge more effectively, coordinate large projects, specialize where appropriate, and delegate responsibilities to the actors and institutions best equipped to perform them.
This would not merely make existing activities more efficient. It would expand the range of things humanity is capable of accomplishing.
Some scientific, technological, infrastructural, and humanitarian projects require such extensive resources and coordination that they may remain impractical for individual states. By combining humanity’s capacities, global institutions could pursue goals that no smaller political community could achieve alone.
2. Governing Advanced AI
Advanced AI may become one of humanity’s most important productive, scientific, military, administrative, and political capacities.
Its development cannot safely be treated as a matter concerning only the states or companies creating it. A relatively small number of actors developing advanced AI recklessly could impose risks on everyone.
Competition also creates incentives for states and companies to accept greater collective risks in pursuit of relative strategic advantages. Even actors that recognize the dangers may fear slowing down if their rivals continue.
Common global rules, monitoring, and enforcement could reduce this pressure and allow humanity to prioritize absolute progress and long-term security over the relative advantage of particular states or organizations.
If advanced AI becomes central to production, research, coordination, and governance, its benefits and control should not be permanently concentrated within a small number of private or national institutions. It should ultimately serve humanity as a whole.
3. Redirecting Resources Away from Conflict
Humanity currently devotes enormous resources to preparing for conflict with itself.
Military competition, deterrence, intelligence operations, weapons development, fortified borders, sanctions, strategic duplication, war, and post-war reconstruction consume labor, material, scientific expertise, and political attention.
The human cost of conflict is therefore greater than the number of people directly killed or injured.
It also includes the hospitals that are never built, the diseases that are never cured, the infrastructure that is never developed, the discoveries that are never made, and the living standards that never improve because resources are instead devoted to competition and destruction.
Conflict reduces humanity’s capacity by turning the abilities of some people against the abilities of others.
Greater political unity would allow more of humanity’s resources to be directed toward development rather than internal rivalry.
4. Reducing the Security Dilemma
Even states that do not seek war may be driven into military competition.
Because governments cannot be certain of one another’s future intentions, defensive measures taken by one state can appear threatening to others. This produces arms races in which each side may regard its own actions as necessary responses to the insecurity created by everyone else.
Common institutions, transparency, inspections, monitoring, enforceable rules, and shared security structures could reduce this uncertainty.
The more effectively security is guaranteed collectively, the less each society must devote to protecting itself against every other society.
A sufficiently developed common political system could gradually transform security from a competition between nations into a shared responsibility of humanity.
5. Internalizing Costs and Benefits
Many decisions affect people who have little or no influence over how those decisions are made.
Climate change, overfishing, cyberattacks, pandemics, financial instability, tariffs, pollution, and unsafe technological development can produce concentrated benefits for particular actors while spreading costs across borders.
The inverse problem also occurs. Some activities create broad benefits that individual actors cannot fully capture. Vaccination, scientific research, infrastructure, environmental protection, and disease prevention may therefore receive less investment than their overall value to humanity would justify.
Fragmented decision-making encourages actors to consider only the costs and benefits that fall within their own jurisdiction or organization.
Global institutions could make it easier to account for everyone affected, aligning decisions more closely with their total consequences for humanity.
They could also help prevent tragedies of the commons, in which individually advantageous behavior collectively destroys a resource or system on which everyone depends.
6. Allocating Resources More Effectively
Human needs, natural resources, productive capacity, infrastructure, and expertise are distributed unevenly across the world.
Political borders do not necessarily correspond to where resources are most urgently needed or where they could generate the greatest benefit.
Under a fragmented system, one society may possess unused productive capacity while another lacks essential goods. Multiple states may invest in redundant projects while important shared needs remain neglected.
Greater global coordination could make it possible to allocate at least some resources according to need, capacity, long-term value, and humanity’s shared priorities rather than treating national boundaries as the ultimate limits of political responsibility.
This would not require every economic decision to be made centrally. Local knowledge, decentralized experimentation, markets, and regional autonomy could remain important. The purpose of global coordination would be to address cases in which fragmented allocation leaves humanity collectively poorer or less capable.
7. Supporting Stable Development
Humanity loses potential whenever people are prevented from developing and contributing their abilities by war, poverty, institutional collapse, insecurity, or lack of access to education and infrastructure.
Well-designed global institutions could help stabilize fragile or developing societies through security guarantees, institutional assistance, financing, knowledge transfer, infrastructure, and impartial mediation.
This should not be understood only as charity from stronger societies to weaker ones.
A society that becomes safer, healthier, better educated, and more productive gains a greater capacity to contribute to science, culture, innovation, production, and the solution of shared problems.
Helping societies develop therefore expands the capabilities of humanity as a whole.
8. Mutual Interdependence Encourages Peace
Greater cooperation and specialization make societies more dependent on one another’s knowledge, production, infrastructure, institutions, and resources.
This raises the cost of conflict.
War no longer destroys only lives and physical infrastructure. It also disrupts supply chains, scientific collaboration, technological development, communication networks, economic systems, and specialized functions upon which all participants depend.
When societies benefit from one another’s continued stability and prosperity, they gain stronger incentives to preserve peaceful relationships.
Interdependence is most likely to support peace when it is mutual, reliable, and protected by credible common institutions. Unequal dependence can instead become a source of coercion, vulnerability, or fear.
Global governance could preserve the productive benefits of specialization while reducing the risk that essential dependencies are weaponized.
9. Tackling Challenges Beyond the Capacity of Smaller Actors
Some challenges exceed the resources or jurisdiction of individual states, particularly smaller and poorer ones.
These include pandemic preparedness, climate change, human trafficking, organized crime, asteroid defense, large-scale space exploration, frontier scientific research, and the safe governance of advanced AI.
But the broader principle extends beyond emergencies.
Humanity can attempt larger and more ambitious projects when it combines its resources. A more coordinated civilization could invest in scientific and technological goals whose costs, risks, and benefits extend across generations and regions.
Global cooperation would therefore allow humanity not merely to survive shared threats, but to pursue possibilities that political fragmentation currently places beyond our collective reach.
Recognizing the Risks
Concentrating authority at the global level would create serious risks of its own.
A poorly designed global state could become distant, unaccountable, or authoritarian. Because its authority would be global, failures or abuses could also become difficult to escape.
A desirable global political system could therefore require strong constitutional limits, separation of powers, judicial independence, transparency, democratic accountability, regional autonomy, and reliable mechanisms for removing those in power.
Authority would generally remain at the lowest political level capable of handling the issue effectively. Local matters would be governed locally, regional matters regionally, and only genuinely global matters would require global authority.
Global unity does not mean cultural uniformity or the elimination of local self-government. Humanity can coordinate globally while preserving diverse communities, identities, institutions, and ways of life.
The General Principle
Civilization has repeatedly expanded human capacity by allowing larger groups of people to cooperate.
Individuals formed communities. Communities developed larger political institutions. Wider systems of trade, law, infrastructure, education, and knowledge allowed people to specialize and build upon the achievements of strangers across both distance and time.
The development of global cooperation would continue this process.
Humanity is not weakened only when people die or resources are destroyed. It is also weakened whenever mistrust, rivalry, poverty, exclusion, or political fragmentation prevents human abilities from being combined productively.
Likewise, humanity does not become stronger merely by avoiding disaster. It becomes stronger when more people can develop their capacities, share their knowledge, coordinate their efforts, and contribute to common projects.
My question is then:
Would you support a democratic global federation?
If not, is your objection to the goal itself, or to the risks and practical difficulties involved in achieving it?
Why AI Should Be Commonly Owned and Controlled — and How Such Cooperation Could Be Organized
Artificial intelligence may become one of humanity’s most important productive, scientific, political, and administrative capacities. If that is the case, the central question is not only what AI will be able to do, but who will own it, who will control its development and use, and who will benefit from the value it creates.
There are at least five major arguments for some form of common ownership and control of advanced AI.
1. AI has global and cross-border consequences
The effects of AI cannot necessarily be contained within the country, company, or institution that develops or deploys it. AI systems may affect other societies through cyberattacks, military applications, economic disruption, automated surveillance, political manipulation, misinformation, or decisions made within globally connected systems.
It is therefore not enough for some states or companies to manage AI responsibly. Reckless, aggressive, or poorly controlled development in one place may create consequences elsewhere.
This makes AI a collective-action problem. Every actor is affected by how others develop and use the technology, while no actor can independently control the entire system. If the risks and consequences of AI are global, then its control must also, to some extent, be organized globally. We cannot adequately control AI separately when its effects do not respect national or institutional borders.
2. AI depends on globally accumulated knowledge and may be developed more efficiently through cooperation
AI is not created in isolation by a single company or group of researchers. It is built on globally accumulated science, publicly funded research, education systems, technological infrastructure, data, energy, hardware, and the intellectual and cultural output of large parts of humanity.
Even when a particular company produces a model, that model rests on a much broader social and international foundation. This gives us reason to view AI not merely as a private product, but as the result of a collective and globally distributed process of knowledge creation.
Large-scale cooperation may also make AI development more efficient. In a fragmented system, competing actors repeatedly solve similar problems, construct parallel infrastructure, conceal discoveries, and duplicate research. Shared research environments, coordinated projects, common computing resources, and the exchange of safety findings could reduce unnecessary duplication.
Such cooperation could also provide a clearer overview of the technology’s development. Public institutions would have a better understanding of which capabilities exist, where the field is heading, what risks are emerging, and which safety measures are required.
3. Common ownership and control can prevent extreme concentrations of power
Advanced AI may provide control over production, research, information, communication, surveillance, coordination, and decision-making. Actors that control the most capable systems may therefore acquire extraordinary economic and political power.
This problem may become more serious as companies and political elites become less dependent on the wider population’s labour. Historically, workers and citizens have possessed some bargaining power because powerful institutions have depended on their labour, knowledge, consumption, cooperation, and participation.
If AI can perform an increasing number of human functions, that dependence may weaken. A small number of actors could then control much of society’s productive and administrative capacity without remaining equally dependent on the majority of the population.
This could produce a historically unusual form of concentrated power: an elite with control over production, knowledge, security, and coordination, but with a declining need for the labour or consent of others.
Common ownership and democratic or international control could help protect broader interests against such domination. The objective does not have to be that every individual directly participates in every technical decision. The objective is that the ultimate control of this capacity should be accountable to a much broader public than the small group that might otherwise own it.
4. Automation is not the problem; the ownership of its benefits is
The replacement of human labour by AI does not have to be harmful. Automation can free people from dangerous, exhausting, repetitive, or time-consuming work while increasing society’s total productive capacity.
The problem arises when people lose their jobs and incomes while the owners of AI systems privately capture the value that those systems produce.
AI does not meaningfully “steal” people’s jobs if the benefits of automation still belong to the people whose labour is no longer required. In that case, automation allows society to maintain or improve living standards while reducing the amount of work that people are forced to perform.
The situation is very different if people collectively contribute to the social, scientific, and economic foundations that make AI possible, are then replaced by the technology, and are finally excluded from its returns. Under those conditions, AI becomes an instrument of dispossession rather than liberation.
The decisive question is therefore not simply how many jobs AI will replace. It is who owns the systems and who receives the value they create.
5. Common ownership and transparency can reduce distrust, arms races, and destructive competition for relative advantage
When advanced AI is controlled by competing states or corporations, it becomes more than a tool for creating value. It also becomes a means of gaining power over others.
Each actor may fear that a competitor will develop a decisive capability first. Even actors that would prefer slower and safer development may therefore feel compelled to accelerate, conceal their progress, lower safety standards, and accept greater risks.
This creates a security dilemma. One actor’s attempt to increase its own security through greater AI capacity may be perceived as a threat by others, who respond by intensifying their own efforts. The result may be an arms race that no actor truly wants, but that no actor feels safe abandoning unilaterally.
The competition is not only about creating benefits in absolute terms. It is also about relative advantage. An actor may oppose the spread of a beneficial technology because it would strengthen rivals, even if its wider use would increase total human welfare.
Common ownership, shared control, and substantial transparency could alter these incentives. If major actors have insight into the development process, share its results, and know that no single actor can monopolize the most advanced capabilities, the value of gaining a secret lead is reduced.
This could shift AI development away from a destructive struggle for relative superiority and toward the expansion of common capacity and shared benefits.
Transparency alone, however, is insufficient. If actors still compete for exclusive control, greater transparency may simply reveal who is falling behind and intensify the race. Transparency must therefore be combined with institutions that share power, risks, technological capacity, and returns.
The Institutional Question
If these arguments are convincing, the next and more difficult question is:
How can sufficiently broad cooperation on AI be achieved, and what should such cooperation look like?
It is not enough to say that AI should be commonly controlled. We must identify the institutions capable of gathering the capital, knowledge, infrastructure, legitimacy, and organizational capacity needed to finance, develop, and govern advanced AI.
A Possible Proposal: Public and International AI Cooperation
My starting point is that large-scale AI cooperation could be organized through states, intergovernmental institutions, or a dedicated global organization for AI.
Individual citizens and smaller organizations do not possess the resources required to finance the computing infrastructure, energy supply, hardware, research environments, data systems, and security structures needed for the most advanced AI development.
Through public budgets, however, humanity already controls more than enough capital to finance such development. The primary limitation is therefore probably not a lack of total resources, but a lack of coordination, trust, political commitment, and suitable institutions.
The necessary resources already exist. They are simply divided among states and are often directed toward competing national or private projects.
Participating states could contribute some combination of:
These contributions could be pooled within a jointly governed organization tasked with developing advanced AI for shared purposes.
This would give public institutions genuine technological and productive capacity rather than leaving them in the position of merely regulating privately controlled systems. Regulation remains necessary, but regulation alone allows private companies to retain the technical expertise, infrastructure, initiative, and ownership of the technology.
Public or international ownership would allow states and societies to become active participants in AI development instead of merely reacting to decisions made by private actors.
The cooperation would not necessarily need to include the entire world from the beginning. It could begin as a coalition of willing states and gradually expand as trust, capacity, and institutional experience develop.
However, the organization must not simply become an instrument of its largest contributors. If influence is determined entirely by financial contributions, technological strength, or military power, the institution may reproduce the same concentration of control that it is intended to prevent.
Its governance would therefore need to address several questions:
The objective should be an institution capable of concentrating enough resources to develop advanced AI while remaining aligned with with shared human interests.
The fundamental issue is therefore not whether humanity can afford to finance commonly controlled AI. Humanity already possesses the required capital, knowledge, and productive capacity.
The real question is how we can create institutions capable of combining those resources, building sufficient trust, and placing the resulting technology under forms of control that serve a broader range of human interests.
Concluding Thesis And Question
The more AI develops into a general capacity for production, knowledge, coordination, and governance, the less desirable it becomes for its development, use, and returns to remain under the exclusive control of a small number of private or national actors.
Common ownership and control can be legitimised as AI has global consequences, is built on global contributions, risks creating extreme concentrations of power, may separate people from the benefits of automation, and can generate distrust and dangerous races for relative advantage.
The question is therefore not only how AI should be governed once it becomes sufficiently powerful. It is also how people can organize early enough to shape who owns it, who controls it, and whose interests it ultimately serves.
So, how can we mobilise people, states, and institutions around the goal of shared ownership and democratic control of AI?
Why AI Should Be Commonly Owned and Controlled — and How Such Cooperation Could Be Organized
Artificial intelligence may become one of humanity’s most important productive, scientific, political, and administrative capacities. If that is the case, the central question is not only what AI will be able to do, but who will own it, who will control its development and use, and who will benefit from the value it creates.
There are at least five major arguments for some form of common ownership and control of advanced AI.
1. AI has global and cross-border consequences
The effects of AI cannot necessarily be contained within the country, company, or institution that develops or deploys it. AI systems may affect other societies through cyberattacks, military applications, economic disruption, automated surveillance, political manipulation, misinformation, or decisions made within globally connected systems.
It is therefore not enough for some states or companies to manage AI responsibly. Reckless, aggressive, or poorly controlled development in one place may create consequences elsewhere.
This makes AI a collective-action problem. Every actor is affected by how others develop and use the technology, while no actor can independently control the entire system. If the risks and consequences of AI are global, then its control must also, to some extent, be organized globally. We cannot adequately control AI separately when its effects do not respect national or institutional borders.
2. AI depends on globally accumulated knowledge and may be developed more efficiently through cooperation
AI is not created in isolation by a single company or group of researchers. It is built on globally accumulated science, publicly funded research, education systems, technological infrastructure, data, energy, hardware, and the intellectual and cultural output of large parts of humanity.
Even when a particular company produces a model, that model rests on a much broader social and international foundation. This gives us reason to view AI not merely as a private product, but as the result of a collective and globally distributed process of knowledge creation.
Large-scale cooperation may also make AI development more efficient. In a fragmented system, competing actors repeatedly solve similar problems, construct parallel infrastructure, conceal discoveries, and duplicate research. Shared research environments, coordinated projects, common computing resources, and the exchange of safety findings could reduce unnecessary duplication.
Such cooperation could also provide a clearer overview of the technology’s development. Public institutions would have a better understanding of which capabilities exist, where the field is heading, what risks are emerging, and which safety measures are required.
3. Common ownership and control can prevent extreme concentrations of power
Advanced AI may provide control over production, research, information, communication, surveillance, coordination, and decision-making. Actors that control the most capable systems may therefore acquire extraordinary economic and political power.
This problem may become more serious as companies and political elites become less dependent on the wider population’s labour. Historically, workers and citizens have possessed some bargaining power because powerful institutions have depended on their labour, knowledge, consumption, cooperation, and participation.
If AI can perform an increasing number of human functions, that dependence may weaken. A small number of actors could then control much of society’s productive and administrative capacity without remaining equally dependent on the majority of the population.
This could produce a historically unusual form of concentrated power: an elite with control over production, knowledge, security, and coordination, but with a declining need for the labour or consent of others.
Common ownership and democratic or international control could help protect broader interests against such domination. The objective does not have to be that every individual directly participates in every technical decision. The objective is that the ultimate control of this capacity should be accountable to a much broader public than the small group that might otherwise own it.
4. Automation is not the problem; the ownership of its benefits is
The replacement of human labour by AI does not have to be harmful. Automation can free people from dangerous, exhausting, repetitive, or time-consuming work while increasing society’s total productive capacity.
The problem arises when people lose their jobs and incomes while the owners of AI systems privately capture the value that those systems produce.
AI does not meaningfully “steal” people’s jobs if the benefits of automation still belong to the people whose labour is no longer required. In that case, automation allows society to maintain or improve living standards while reducing the amount of work that people are forced to perform.
The situation is very different if people collectively contribute to the social, scientific, and economic foundations that make AI possible, are then replaced by the technology, and are finally excluded from its returns. Under those conditions, AI becomes an instrument of dispossession rather than liberation.
The decisive question is therefore not simply how many jobs AI will replace. It is who owns the systems and who receives the value they create.
5. Common ownership and transparency can reduce distrust, arms races, and destructive competition for relative advantage
When advanced AI is controlled by competing states or corporations, it becomes more than a tool for creating value. It also becomes a means of gaining power over others.
Each actor may fear that a competitor will develop a decisive capability first. Even actors that would prefer slower and safer development may therefore feel compelled to accelerate, conceal their progress, lower safety standards, and accept greater risks.
This creates a security dilemma. One actor’s attempt to increase its own security through greater AI capacity may be perceived as a threat by others, who respond by intensifying their own efforts. The result may be an arms race that no actor truly wants, but that no actor feels safe abandoning unilaterally.
The competition is not only about creating benefits in absolute terms. It is also about relative advantage. An actor may oppose the spread of a beneficial technology because it would strengthen rivals, even if its wider use would increase total human welfare.
Common ownership, shared control, and substantial transparency could alter these incentives. If major actors have insight into the development process, share its results, and know that no single actor can monopolize the most advanced capabilities, the value of gaining a secret lead is reduced.
This could shift AI development away from a destructive struggle for relative superiority and toward the expansion of common capacity and shared benefits.
Transparency alone, however, is insufficient. If actors still compete for exclusive control, greater transparency may simply reveal who is falling behind and intensify the race. Transparency must therefore be combined with institutions that share power, risks, technological capacity, and returns.
The Institutional Question
If these arguments are convincing, the next and more difficult question is:
How can sufficiently broad cooperation on AI be achieved, and what should such cooperation look like?
It is not enough to say that AI should be commonly controlled. We must identify the institutions capable of gathering the capital, knowledge, infrastructure, legitimacy, and organizational capacity needed to finance, develop, and govern advanced AI.
A Possible Proposal: Public and International AI Cooperation
My starting point is that large-scale AI cooperation could be organized through states, intergovernmental institutions, or a dedicated global organization for AI.
Individual citizens and smaller organizations do not possess the resources required to finance the computing infrastructure, energy supply, hardware, research environments, data systems, and security structures needed for the most advanced AI development.
Through public budgets, however, humanity already controls more than enough capital to finance such development. The primary limitation is therefore probably not a lack of total resources, but a lack of coordination, trust, political commitment, and suitable institutions.
The necessary resources already exist. They are simply divided among states and are often directed toward competing national or private projects.
Participating states could contribute some combination of:
These contributions could be pooled within a jointly governed organization tasked with developing advanced AI for shared purposes.
This would give public institutions genuine technological and productive capacity rather than leaving them in the position of merely regulating privately controlled systems. Regulation remains necessary, but regulation alone allows private companies to retain the technical expertise, infrastructure, initiative, and ownership of the technology.
Public or international ownership would allow states and societies to become active participants in AI development instead of merely reacting to decisions made by private actors.
The cooperation would not necessarily need to include the entire world from the beginning. It could begin as a coalition of willing states and gradually expand as trust, capacity, and institutional experience develop.
However, the organization must not simply become an instrument of its largest contributors. If influence is determined entirely by financial contributions, technological strength, or military power, the institution may reproduce the same concentration of control that it is intended to prevent.
Its governance would therefore need to address several questions:
The objective should be an institution capable of concentrating enough resources to develop advanced AI while remaining aligned with with shared human interests.
The fundamental issue is therefore not whether humanity can afford to finance commonly controlled AI. Humanity already possesses the required capital, knowledge, and productive capacity.
The real question is how we can create institutions capable of combining those resources, building sufficient trust, and placing the resulting technology under forms of control that serve a broader range of human interests.
Concluding Thesis And Question
The more AI develops into a general capacity for production, knowledge, coordination, and governance, the less desirable it becomes for its development, use, and returns to remain under the exclusive control of a small number of private or national actors.
Common ownership and control can be legitimised as AI has global consequences, is built on global contributions, risks creating extreme concentrations of power, may separate people from the benefits of automation, and can generate distrust and dangerous races for relative advantage.
The question is therefore not only how AI should be governed once it becomes sufficiently powerful. It is also how people can organize early enough to shape who owns it, who controls it, and whose interests it ultimately serves.
So, how can we mobilise people, states, and institutions around the goal of shared ownership and democratic control of AI?
Capitalism has been effective because it channels strong human motivation into production and innovation. Communism and parts of the political left have often underestimated how dependent social development has been on such incentives.
AI, however, changes this relationship. As innovation and production become increasingly automated, society no longer needs to rely to the same extent on competition, material desperation, and unequal rewards in order to remain efficient.
Common ownership and egalitarian distribution may therefore shift from being developmentally costly ideals to becoming rational and effective strategies for giving humanity access to automated capacity.
The challenge is to design the transition so that human motivation is preserved wherever it is still needed, while dependence on human labor and the need for private incentives decline in step with AI’s growing capabilities.
How might such a design look like? How can motivation sufficient enough to drive progress be preserved, whilst less and less is required of us to sustain our lives.
Capitalism has been effective because it channels strong human motivation into production and innovation. Communism and parts of the political left have often underestimated how dependent social development has been on such incentives.
AI, however, changes this relationship. As innovation and production become increasingly automated, society no longer needs to rely to the same extent on competition, material desperation, and unequal rewards in order to remain efficient.
Common ownership and egalitarian distribution may therefore shift from being developmentally costly ideals to becoming rational and effective strategies for giving humanity access to automated capacity.
The challenge is to design the transition so that human motivation is preserved wherever it is still needed, while dependence on human labor and the need for private incentives decline in step with AI’s growing capabilities.
Value does not exist independently of someone who values.
Nothing is inherently good or bad. A technology, institution, or economic development has value only in relation to the wills it serves or obstructs.
By will, I mean an individual’s goals, interests, preferences, priorities, and convictions.
Will alone does not shape reality. It requires capacity: knowledge, labor, resources, technology, organization, and influence.
Because people’s wills differ, they may become allies or opponents. Where their aims overlap, they can coordinate. Where they conflict, they compete over which will shapes reality.
Coordination allows people to combine their capacities and achieve more than they could alone. It requires compromise, but the capacity gained through cooperation can outweigh what each participant gives up.
Politics is therefore both coordination and conflict over how coordination is structured.
Different structures serve different wills. A small but powerful group can organize vast collective capacity around its own interests, even when most people would prefer a different outcome.
This is the central political issue surrounding AI.
AI is not inherently good or bad. It is a powerful amplifier of capacity. How that will come to affect us depends on who controls it and whose wills it serves.
The danger is not simply AI. It is a future in which a few corporations control advanced AI, become less dependent on everyone else, and gain the capacity to shape society around their own interests.
The rational response is not merely to condemn them or reject the technology. It is to organize competing capacity and fight for control over how AI is developed, owned, and applied.
Most people share at least some basic aims: survival, the wellbeing of those they love, and the possibility of living comfortably. Those overlapping wills create a basis for cooperation.
The real questions are not:
Is AI good or bad?
Who is morally right?
They are:
Who will control AI?
Whose wills will it serve?
And who will organize enough capacity to shape the answer?
Democratic Control of AI — Before AI Becomes Society’s Primary Productive Force
I see artificial intelligence as one of humanity’s greatest opportunities. My goal is not to halt its development, but to build the institutions that allow us to control it in a way that benefits all of humanity.
The central question, therefore, is not whether AI is good or bad.
The central question is:
Who will have the power to decide how AI is used when it becomes one of society’s most important productive and governing resources?
From Human Labor to Intelligence as the Primary Productive Force
Throughout history, human labor has been the primary productive force. Because of this, people have possessed economic value that has given them influence over the development of society.
AI changes this.
As AI and automation continue to advance, human labor may gradually lose its unique economic role. AI does not need to become perfect or replace every human task. It only needs to become sufficiently capable and significantly less expensive across an expanding range of activities for economic incentives to drive widespread automation.
This is not a certainty. It is, however, a plausible enough trajectory that we should begin preparing our institutions today.
The Real Question Is About Power
Public debate on AI often focuses on employment, misinformation, safety, or regulation.
These are important issues.
But I believe they obscure a more fundamental question.
If AI becomes the dominant resource for production, coordination, scientific discovery, and possibly even decision-making, then control over AI becomes control over a substantial part of society’s future development.
The defining political question is therefore not what AI is capable of doing.
The defining political question is who decides what AI should be used for.
Democracy Depends on Mutual Dependence
Historically, economic and political power have been constrained by mutual dependence.
Businesses needed workers.
Governments needed citizens.
Capital needed labor.
No single actor could become completely independent of the rest of society.
If AI and automation dramatically reduce these mutual dependencies, the balance of power may fundamentally change. Those who control the most powerful AI infrastructure could acquire a level of influence with little historical precedent.
This is not an argument against AI.
It is an argument that democracy must evolve alongside technology.
The Goal Is Not to Preserve Work in of itself
I am not arguing that people should continue working simply for the sake of having work.
If a task can be automated in a way that frees human time and creates greater prosperity, we should generally welcome it.
Deliberately preserving inefficient work merely to keep people occupied risks creating artificial employment instead of genuine opportunity.
Human value does not lie in being inexpensive labor.
Human value lies in being the people whose lives and interests society ultimately exists to serve.
Democratic Governance Before Uncontrolled Acceleration
I do not want to stop AI development.
I want to build the institutional foundations that allow us to develop it with confidence.
This means that we should already be discussing and developing:
This does not necessarily mean that governments should own all AI.
It means that no single actor—whether a corporation, a government, or a small economic elite—should be able to unilaterally control a technology capable of shaping humanity’s future.
A Call to Action
AI has the potential to create the greatest increase in prosperity in human history.
It can free humanity from much of the labor currently required for survival.
It can accelerate science, medicine, education, and technological progress on an unprecedented scale.
Precisely because of this potential, the question of control becomes essential.
The greater a technology’s capacity to shape society, the broader the democratic influence over its goals and application should be.
My goal is therefore not to slow AI.
My goal is to help organize people around a simple principle:
The most powerful technology humanity has ever created should ultimately be developed and governed in humanity’s shared interest—not controlled by a small number of powerful actors.
Democratic Control of AI — Before AI Becomes Society’s Primary Productive Force
I see artificial intelligence as one of humanity’s greatest opportunities. My goal is not to halt its development, but to build the institutions that allow us to control it in a way that benefits all of humanity.
The central question, therefore, is not whether AI is good or bad.
The central question is:
Who will have the power to decide how AI is used when it becomes one of society’s most important productive and governing resources?
From Human Labor to Intelligence as the Primary Productive Force
Throughout history, human labor has been the primary productive force. Because of this, people have possessed economic value that has given them influence over the development of society.
AI changes this.
As AI and automation continue to advance, human labor may gradually lose its unique economic role. AI does not need to become perfect or replace every human task. It only needs to become sufficiently capable and significantly less expensive across an expanding range of activities for economic incentives to drive widespread automation.
This is not a certainty. It is, however, a plausible enough trajectory that we should begin preparing our institutions today.
The Real Question Is About Power
Public debate on AI often focuses on employment, misinformation, safety, or regulation.
These are important issues.
But I believe they obscure a more fundamental question.
If AI becomes the dominant resource for production, coordination, scientific discovery, and possibly even decision-making, then control over AI becomes control over a substantial part of society’s future development.
The defining political question is therefore not what AI is capable of doing.
The defining political question is who decides what AI should be used for.
Democracy Depends on Mutual Dependence
Historically, economic and political power have been constrained by mutual dependence.
Businesses needed workers.
Governments needed citizens.
Capital needed labor.
No single actor could become completely independent of the rest of society.
If AI and automation dramatically reduce these mutual dependencies, the balance of power may fundamentally change. Those who control the most powerful AI infrastructure could acquire a level of influence with little historical precedent.
This is not an argument against AI.
It is an argument that democracy must evolve alongside technology.
The Goal Is Not to Preserve Work in of itself
I am not arguing that people should continue working simply for the sake of having work.
If a task can be automated in a way that frees human time and creates greater prosperity, we should generally welcome it.
Deliberately preserving inefficient work merely to keep people occupied risks creating artificial employment instead of genuine opportunity.
Human value does not lie in being inexpensive labor.
Human value lies in being the people whose lives and interests society ultimately exists to serve.
Democratic Governance Before Uncontrolled Acceleration
I do not want to stop AI development.
I want to build the institutional foundations that allow us to develop it with confidence.
This means that we should already be discussing and developing:
This does not necessarily mean that governments should own all AI.
It means that no single actor—whether a corporation, a government, or a small economic elite—should be able to unilaterally control a technology capable of shaping humanity’s future.
A Call to Action
AI has the potential to create the greatest increase in prosperity in human history.
It can free humanity from much of the labor currently required for survival.
It can accelerate science, medicine, education, and technological progress on an unprecedented scale.
Precisely because of this potential, the question of control becomes essential.
The greater a technology’s capacity to shape society, the broader the democratic influence over its goals and application should be.
My goal is therefore not to slow AI.
My goal is to help organize people around a simple principle:
The most powerful technology humanity has ever created should ultimately be developed and governed in humanity’s shared interest—not controlled by a small number of powerful actors.