Frequently asked questions.

You are building the most consequential technology in human history. If frontier AI fulfils even a fraction of its promise, it may create wealth on a scale never before seen — not merely billions, not merely trillions. Global output could rise tenfold, and perhaps far more.

But a richer world is not automatically a fairer world. If the value of AI is captured only by founders, employees, investors and shareholders, the result will not be abundance for humanity — it will be concentration on a scale no society has ever survived intact.

We are writing now because the settlement must be made before the new economy is locked in.

Frontier AI is not being built from nothing. It is built from humanity's writing, code, art, conversation, music, science, medicine, law and history — the accumulated work of every generation that came before us. It is trained on a corpus that belongs, in any honest accounting, to humanity as a whole.

No company created that inheritance alone. No investor funded it alone. No nation owns it alone.

If AI is trained on humanity, humanity deserves equity.

We are clear-eyed about what is coming for human work. If AI can do the work of doctors, lawyers, teachers, analysts, engineers, designers, administrators, drivers, factory workers and carers, then billions of people may soon have no reliable way to exchange their labour, skill or knowledge for income.

The prosperity may be real. The productive capacity may be astonishing. But unless people are taken along, that prosperity will arrive as fear, insecurity and dispossession.

The answer is not to stop AI. The answer is to make every human an owner in the age AI creates.

Every living human, without exception, will receive one equal, permanent, non-transferable share. It can never be sold, traded, inherited, accumulated, pledged, mortgaged or captured by governments, corporations, lawyers or financial intermediaries. It exists because the person exists. On birth, a new share begins; on death, the share ends.

Each year, that share produces an annual compute dividend. The dividend, once issued, belongs to the person — adults with capacity may use it, sell it, gift it, donate it, delegate it, transfer it or let it lapse as they choose.

For children under 18 and people lacking capacity, the dividend still belongs to them. It is not owned by parents, guardians, governments or institutions; it must be protected and used only in their interests until they are able to control it themselves.

One human. One share.

We are not asking you to slow down. We are not asking governments to confiscate. We ask that OpenAI and Anthropic each place twenty-five per cent of fully diluted equity into an irrevocable global trust for living humanity.

The same commitment should bind any future frontier AI company of comparable global importance, including any successor that displaces OpenAI or Anthropic through a future technological breakthrough.

Each year, twenty-five per cent of distributable earnings should be converted into compute at audited cost — compute is the native currency of the AI age — and divided equally among all living humans. The share can never be sold. The dividend is theirs.

This declaration is not anti-AI. It is not anti-business. It is not a party-political demand, a welfare programme, or a call for confiscation. It is the social licence for the AI age.

We believe frontier AI can create extraordinary abundance — and that abundance must include the people whose knowledge, labour, culture and data made it possible. No border, government, company or bureaucracy should stand between a human being and their equal share in the age of artificial intelligence.

One human. One share. If AI is trained on humanity, humanity deserves equity.

No one can know the future value of frontier AI, but we should be honest about the scale of what is being built. The global economy is roughly $125 trillion a year, much of it from human knowledge and physical work — analysis, design, administration, logistics, care, teaching, engineering, law, medicine, manufacturing and transport.

If advanced AI agents and robots can perform much of that work, the cost of labour could fall towards zero in many sectors, even as total output rises dramatically. A tenfold increase sounds radical; a hundredfold sounds impossible — but this is not a normal industrial transition.

If global output rose one hundredfold, a $100 trillion labour-and-knowledge economy could become a $10,000 trillion economy. If the dominant frontier AI companies captured that value at a 50% profit margin, that is roughly $5,000 trillion of annual profit. At a price-to-earnings multiple of 20, that implies a combined valuation near $100,000 trillion — $100 quadrillion.

A twenty-five per cent humanity share of that would be worth around $25 quadrillion. Divided equally among 8.3 billion people, that is about $3 million per person in capital value — and at a 5% annual yield, an annual compute dividend of about $150,000 per person per year.

These numbers are not a promise. They are a scale check. If the people building frontier AI are right about what it can do, then One Human One Share is not symbolic — it is the difference between a future where billions are made economically unnecessary, and one where every human owns a permanent share in the abundance their civilisation helped create.

The share can never be sold. The dividend is theirs.

The share belongs to the person. It does not depend on nationality, alliance, ideology, sanctions, borders, citizenship, or the approval of any state. Every living human should receive one equal, permanent, non-transferable share, because the claim is based on being human, not on being born in the right country.

Some governments may try to block access to frontier AI — restricting compute, censoring models, or preventing their citizens from receiving the full benefit of the dividend. That may happen. But it will not be sustainable forever. A country that cuts its people off from advanced intelligence, automation, education, translation, science and productivity will fall behind those that do not.

The entitlement should therefore remain waiting for them. If a person cannot access their compute dividend because of national restrictions, conflict, sanctions or censorship, their share should not disappear — it should remain theirs, and when access becomes possible, the benefit should be available to them.

One Human One Share does not ask AI companies to serve governments. It asks them to recognise humans.