I sought to lead the World Health Organization (WHO) because the rules of intelligence are becoming the rules of health. My name will not be on the ballot. The responsibility remains. I will now work to organise the political and economic power needed to put intelligence at the service of humanity – and to keep it from turning against us.

Nominations for the next WHO Director-General closed on 24 September. Seven names are on the list. Mine is not. Those who received me as a candidate this summer deserve more than a sentence. The short version is that I chose a task I judged more urgent than the leadership of any single institution. The longer version is this essay, because the vision was never about me, and the question at its centre still awaits an answer from any of the six.

The question is simple to state, and it has been asked of every technology that has ever entered the clinic. The global health architecture is about to acquire a new layer of intelligence. For whom will it work, and under whose control? And, more urgently now: who will keep it from turning against us?

What the field taught me: Social, political and intelligence determinants of health

I came to that question slowly, through two kinds of work that rarely speak to each other.

More than 20 years ago, my first humanitarian posting as a physician was at Hanawon, the centre where people who have escaped North Korea spend their first months in the South. 

For three years I examined people who had crossed borders on foot and spent years in hiding. They arrived with tuberculosis and trauma, and with structural violence that had settled into the body. 

But my clinic was the end of their illness’s story, not its beginning. The social conditions that had made them ill lay years upstream: a famine that was political before it was agricultural, borders that turned flight into a crime, a stateless existence in which seeing a doctor meant risking deportation. What I treated were the sediments of a social order – the social determinants of health, arriving one body at a time.

With Médecins Sans Frontières, I followed those conditions back to their source: the borderlands people had fled, and later, with other humanitarian colleagues, to the Kashmir earthquake, the jungles of Papua New Guinea, and the refugee camps of the Democratic Republic of Congo. 

There I learned a lesson the textbooks phrase too politely. Illness in a crisis is rarely an accident of biology. It is the body’s registration of decisions made elsewhere. A woman denied even refugee status turns to a trafficker not because she misjudges the risk, but because every legal door has closed and return is the one certainty she cannot survive. The well-meaning people in white vehicles – I was one of them – too often arrive without the history that made the poverty they came to relieve, mistaking colonial residue for local misfortune.

Twenty years later, the citizens of Osan, a working city south of Seoul, sent me to the National Assembly, and there I learned the other half of the story. I watched the grief of Itaewon, where 159 people died in a single night, bent to political need and then erased. I watched the martial law decree of 3 December 2024 outlaw the Assembly in one clause and order striking doctors back to their wards in another – one text aimed at legislators like me and physicians like me.

And on the presidential campaign and the National Planning Committee, among the few who write national strategy, I learned how a line in my policy book becomes a ministry – or quietly a footnote no one answers for. 

Behind every failure of health care I had witnessed stood not a shortage but a choice: a political determinant. Power decides not only who is treated, but who is recognised as ill, and who has the power to judge whether or what treatment is provided.

‘Intelligence’ – the determinant behind the determinants

Between the social and political determinants, a third took shape, and it ran deeper than either. The resource on which health systems have depended longest, and which they have distributed most unequally, is neither medicine nor money. It is human judgement: the capacity to recognise who is ill, to decide what should be done, and to act. I call the conditions that govern the social distribution of that judgement the intelligence determinants of health. 

Across much of the world, a single nurse makes the decisions on which 10,000 lives may depend, with no one to consult and no second opinion within a day’s travel. Economists, scientists and policy experts are distributed across the world just as unevenly, and that distribution may itself be a determinant of the social and political conditions I had spent 20 years treating – the determinant beneath the determinants.

For the first time, a machine can carry a part of that judgement to places it has never reached. Governments and companies will decide within the next few years whether it reaches that nurse or only the hospitals that can pay – and they will decide much of it in private contracts whose terms no one outside the signing room will ever read: an old pattern in a new form, the determinants of health settled where the people they determine cannot see.

What I stood for: Climate, Health, and AI

I wanted WHO to write the rules for this emerging layer of intelligence before the market did. In an emergency, the interval between a need and a standard is filled by whoever arrives first. WHO is the one institution every health ministry listens to, but the distance between a principle agreed in Geneva and a contract signed in a ministry is where the future of health care is being settled.

For a decade, since AlphaGo defeated the world’s best Go player,Lee Sedol, born on my mother’s island, I have worked on this new layer of intelligence from the outside: on AI-augmented health systems for low-resource settings, and more recently on Korea’s new societal model, the “AI Universal Basic Society” (AI UBS), which treats intelligence as public infrastructure for essential services, beginning with health care. 

In May, nine UN agencies including WHO, launched the Global AI Hub, a shared operational campus of the UN system, and in September Korea put both proposals before the General Assembly.

But WHO itself was not moving at the speed the moment demanded. Member states cut its base budget for 2026–27 from $5.3 billion to $4.2 billion, and development assistance for health fell to its lowest level in roughly 15 years. An organisation that has lost a fifth of its budget and many of its ablest people does not run towards new fields; it defends the ones it has. Meanwhile the largest technology companies were signing national-scale agreements for health and education, most on undisclosed terms.

So I did what a physician does when a referral is taking too long. I went to the patient myself. In July, I declared.

The core of my policy manifesto, One Doctor, Eight Billion Patients, is a single proposition: Universal Intelligence Coverage (UIC) for health. Universal health coverage asks whether people can receive care without being ruined by the cost. UIC asks whether judgement – human, machine or both – reaches everyone, and who answers when it fails. 

I proposed a “1.5-tier” model of primary health care: frontline workers equipped with AI decision support and supervised remotely by human professionals, with responsibility for diagnosis and treatment kept firmly with people.

Climate finance was the second strand. Less than 1% of it reaches health. I argued that a climate-resilient health system is itself a frontline mechanism of adaptation, and proposed that WHO use the Green Climate Fund accreditation it secured this year to channel roughly $1 billion into primary care without drawing a dollar from its shrinking core budget. 

The larger point was about dependence. Global health has run for decades on voluntary, earmarked money from a handful of governments, and the past two years have shown how quickly that base erodes when one of them turns inward. 

Blended finance is the way out: climate financing anchoring the public health layer, the private capital now pouring into the AI transition financing the intelligence layer above it, and public terms binding both to public goals.

Over 10 weeks I took these ideas to three continents – Africa, Asia and Europe – and to the leaders of their regional organisations. What we built together rests on one principle: regional health sovereignty – a decolonised global health architecture, powered by a new layer of intelligence and a new mechanism of finance. 

In Africa, it is “African AI for African health”, running on African infrastructure and answering to African law. In the Pacific Islands it is the “Pacific AI Initiative”, a regional platform for climate-resilient health systems. All want to build, not merely to buy.

After stepping back

I believe the diagnosis was right and the prescription was ready. I had been nominated by the Parliament as a WHO candidate for DG. But at the last minute, circumstances changed.  

The responsibility now passes to one of the other seven candidates, and I wish each of them well. Whoever becomes Director-General will find in me a partner. But the campaign taught me something I had not fully grasped when it began.

Meeting heads of government, frontier researchers and leaders of AI and biotechnology companies, I heard the same assessment with increasing frequency: the convergence of autonomous AI and biotechnology is entering a stage at which it could generate an existential threat of a new kind: a pandemic made by humans and machines together. 

The frontier laboratories now say so themselves: by their own published assessments, their latest models have reached the thresholds at which they could materially assist biological misuse, and the companies that build them ask publicly to be regulated. A risk I had been warning about for years was arriving far faster than I had expected. The window is closing: capability is advancing by the quarter, public control by the decade.

These are not two separate agendas. Bringing validated intelligence to a rural clinic and keeping unvalidated intelligence from harming the world are two answers to the same question – for whom, and under whose control, does intelligence work? 

But this answer lies beyond WHO’s reach. Responding to a pandemic, whatever its origin, is WHO’s mandate. Governing the roots of this risk – the race in frontier AI and its fusion with biotechnology – belongs to no institution at all. 

As with the nuclear competition of the last century, the world has no mechanism to slow a race centred on the United States and China, and the voice of industry overwhelms that of states. The world has visions and forums enough; it has no institution and no budget line for the AI society that is already arriving.

What it does have is the makings of a coalition. The Global South champions and the middle powers that hold technology and finance can, together, assemble the strength to shape that AI society. I have concluded that this work was more urgent than any position in any organisation. Over these months, leaders serving and former – in governments, international organisations, the media and academia – reached the same conclusion and asked me to help convene such a coalition. I In my own country, the expectation came from the highest level of government. I chose it.

A humanitarian learns early to go not where the doctors already are, but where there are none. The WHO post will be well filled; this work was more pressing, and almost no one was standing in line for it.

What I will do at the existential crossroads of the AI society

Most proposals for AI governance begin with visions and institutions. I have sat in enough of those rooms to know how slowly they fill. Mine begins with a buyer – and in a market of scattered ministries, a buyer first has to be assembled. In a revolution led by private actors, the power a public actor can use fastest is to decide what it buys and on what terms.

Outside the few countries that build the intelligence infrastructure itself, the Global North will mostly use AI to replace human labour in systems that already exist, and substitution alone does not make a new economy.

In the Global South, the need has always outrun the supply–the clinic without a doctor, the classroom without a teacher, the household without a bank–and intelligence that can finally meet it creates a new economy rather than replacing one: a trillion-dollar market in which the buyer, not the supplier, should set the terms–and no one has yet organised a hundred fragmented buyers into one.

That is the work I will do, and the order is the strategy. Ask industry to regulate itself before you hold any market leverage, and the rules it writes will hold you. I will organise a coalition that can negotiate, purchase and legislate, govern–in that order.

First, a Planetary Intelligence Network, solidarity for sovereignty. Sovereignty over intelligence will not be won one ministry at a time; only a network of leaders who can decide can build it. I will convene it within months, not years. Any leader who shares these concerns – sovereignty over external intelligence, existential risk – has a seat. This is a network, not a bloc, and its first effect is simple: ministries that have always faced the same supplier alone begin to negotiate as one.

Second, New 3 by 5, collective purchasing power. Twenty years ago, Dr Lee Jong-wook set the world a target he called 3 by 5: three million people on HIV treatment by 2005. I propose a new 3 by 5: three million AI-augmented primary health care (PHC) workers within five years. We will augment first the community health workers, nurses and clinical officers who provide most first-contact care, raising their clinical and epidemiological capability toward that of a primary care physician, under remote human supervision. 

This is physician-level judgement arriving in three million places that have never had it. Connected across a continent, AI-augmented PHCs become a connected health system and pandemic preparedness–and a single buyer large enough to set terms. If a pandemic made by humans and machines ever begins, its first case will present not to a frontier lab but to one of them. Access and early warning are one investment.

Third, a Public Intelligence Compact, collective regulating power. A buyer of that size can set terms. The Compact is one set of conditions for any system that shapes a decision about a person in a public service: independent validation, data management under national law, continuity of service, the right to audit, transparent pricing. As a declaration, these are pieties; as a procurement rule, they are the law of the market. 

The Compact is neutral as to flag: any company that meets the terms is welcome. Sovereignty means a country can choose, question, change and, when necessary, leave the systems it uses. Its companion I call Firebreak – the line a fire does not cross. We cannot put out the fire of frontier development, but we buy the very models that could start one, and a buyer can set conditions today: no DNA synthesis order unscreened, no model in a public service unevaluated, no incident unreported. 

The Compact is the first line of defence, Firebreak the last, and the three million workers watching for the first spark are the sensors in between.

Fourth, a Public Intelligence Foundation and Existential Risk Council, a blended governance. This step evolves with purchasing power and regulating power. The Foundation is a shared operational platform with the global AI companies that will translate private AI back into public intelligence on the Compact’s terms. 

Beside it, and firewalled from it, stands the Existential Risk Council, independent of the foundation’s money, because a body that depends on industry cannot also judge its dangers. It will make independent evaluation of existential risks such as biological misuse–with incident reporting and synthesis screening–the price of entry for every model and the laboratory that trained it. The council will define itself against no state.

None of this replaces the United Nations. The network brings will, leverage and pace; the UN brings implementation and universality–the Global AI Hub’s agencies beyond health, and WHO for the clinical standard, which we will adopt, not rival. I know how this fails: a foundation captured by its funders, pilots that vanish with the donor cycle, and the oldest failure of all–help that arrives without asking the people it is for.

That is the work I am committing to lead.

Eight billion for one patient.

Jiho Cha is a humanitarian physician, global health scholar and Member of the National Assembly of the Republic of Korea, and co-chairs the Lancet Commission on Sea-Level Rise, Health, and Justice. For two decades, he has worked with marginalised populations in the Middle East, Africa and Asia, and with the displaced people caught between them. He was Professor of Humanitarian Studies at the University of Manchester and of AI and Future Strategy at KAIST, and designed Korea’s AI Universal Basic Society framework and the Global AI Hub, a joint campus of nine UN agencies. His policy manifesto, One Doctor, Eight Billion Patients, was prepared for his candidacy for WHO Director-General. The views expressed are his own.

Image Credits: Jiho Cha, Ron Lach via Pexels.