How intelligence reaches a person
Medicine already knows most of what would let people live long and healthy lives, and the work of the coming decades is carrying that knowledge to each of them, in hospitals, clinics, ministries and homes, until the place a person is born no longer decides how long they live.
A Tuesday in clinic
The case that follows is illustrative.
A woman of 58 comes to a hospital outpatient clinic because the stairs at home have started to leave her breathless. Before she sits down, her doctor has read a short summary of the last three years of her record. Her blood pressure has crept up. A kidney result last spring was borderline. In the autumn her family doctor started her on a new medicine. She agrees to have the consultation recorded, and while she talks he keeps his eyes on her.
By the time she stands to leave, the note is drafted, along with a letter to her family doctor and the codes for the visit. He reads all three, corrects a dose, adds a line about what he will look for on her echocardiogram, and signs. The scan is booked before she reaches the car park. That evening she reads a plain account of the visit in her own language. When the result comes back, it goes to someone who can see her whole story, and she hears from the hospital that week.
Nothing in that visit depends on technology still to be invented. It is simply rare, and the work ahead is to make it ordinary wherever people live.
The goal we have set for Wai and the Wai Foundation is too large for any one company, a longer, healthier and happier life for everyone. In 2024 everyone meant 8.2 billion people. The United Nations expects around 10.3 billion by the mid 2080s 1.
The trend is already with us. Global life expectancy at birth reached 73.3 years in 2024, and by the late 2050s more than half of all deaths are projected to occur at age 80 or older, up from 17 percent in 1995 1. I studied neuroscience, then medicine, then public health, and each taught me a version of the same lesson. Most of what would keep people well for a long life is already known somewhere in medicine. What is missing is getting it, at the standard of an expert clinician, to the person who needs it at the moment they need it.

What the models can do now
In ordinary clinics and wards this year, software of this kind can listen to a consultation and draft the note, the letter and the codes for a clinician to check. It can read a long, scattered history and summarize it well enough to be useful before the patient walks in. It can answer a clinical question and show its sources. At night it can answer a patient's message, book an appointment within the provider's rules or explain a bill, in many languages. Anything clinical still belongs to a clinician, who reads it and signs it.
It is getting better and cheaper fast. The cost of running a model as capable as the one behind the first release of ChatGPT fell more than 280 times between November 2022 and October 2024 2. Researchers at METR found that the length of tasks frontier agents can complete on their own, measured by how long the tasks take skilled people, doubled roughly every seven months over six years 3. In biology the change has already arrived. AlphaFold predicted the structures of virtually all of the 200 million proteins researchers have identified, more than two million people in 190 countries have used it, and its creators shared the 2024 Nobel Prize in Chemistry 4.
Doctors have noticed. In 2023, 38 percent of US physicians told the American Medical Association they used AI in some form in their work 5. In the association's 2026 survey of nearly 1,700 physicians, the figure was 81 percent 6. Among the general population, generative tools reached 53 percent adoption within three years, faster than the personal computer or the internet 7.
METR counts a task as within a model's reach when the model succeeds about half the time. In medicine half the time would be dangerous, which is why a clinician reads and signs, and why these systems have to be tested on real care.
Why it takes time
In 1990 the economic historian Paul David compared the computer to the electric dynamo 8. Factories took up electric power around the turn of the twentieth century, and for years productivity barely moved. The gains came only when owners redesigned the factory itself around the new source of power. Erik Brynjolfsson, Daniel Rock and Chad Syverson later called this a productivity J curve. Organizations first spend on new processes, training and data, measured output stays flat for a while, and then the gains compound 9. Medicine has its own version. Zoë Slote Morris, Steven Wooding and Jonathan Grant examined the familiar estimate that research takes about 17 years to reach routine practice. They showed how much complexity hides inside that number, and still several independent measurements landed near two decades 10. Medicine is seldom short of knowledge. What slows it is the chain that carries knowledge into a consultation, and people who understand each link can shorten it.
I count five links. The first is the record, which in medicine grew up across systems built for billing and storage. In Medicine on one record, Sofian Youssef, who built the Wai platform, explains what changes when it is joined into one. The second is regulation, medicine's long memory of what keeps patients safe, written into standards such as England's DCB0129 for manufacturers of health software 11 and into NHS England's guidance that keeps the clinician responsible for reviewing what ambient scribing products produce 12. The third is trust. Patients and clinicians extend it on evidence, and it grows fastest when the evidence is published. The fourth is the clinician's judgment. Every rollout I have seen succeed was designed around the person whose name goes on the decision. The fifth is money. A new tool has to make sense in its first week to the people who pay for it and the people who use it. Each link gives way to patient work, and when all five are worked on together, the distance between a model and a patient shrinks from decades toward months.
Medicine first
The same chain runs through law, finance, education, logistics and public administration. Each runs on a record, a body of rules, trust and skilled judgment, and as each is rebuilt, expert help that was once scarce becomes something a farmer or a small business owner can call on when they need it.
Medicine is different in one way. It decides the length and shape of a life, whether a child reaches school age, whether a mother meets her grandchildren, whether a person's last twenty years are active ones. That is why Wai began with medicine, and why it works across the whole path of care, for hospitals, health systems, governments, payers, clinics and individuals.
Medicine can move fast when it decides to. In 2008, 9 percent of US nonfederal acute care hospitals had a basic electronic health record, and by 2024, 99 percent had a certified one 13. The record arrived. It has yet to serve the consultation. In one study, family physicians spent about 5.9 hours of an 11.4 hour workday in the record 14. In another, physicians in four specialties spent nearly two hours on the record and desk work for every hour with patients 15.
At Kaiser Permanente in Northern California, 7,260 physicians used an ambient scribe across more than 2.5 million encounters between October 2023 and December 2024, and 88 percent of the physicians who answered a survey said it improved their interactions with patients 16. In a randomized trial at UCLA Health with 238 physicians across 14 specialties, one scribe reduced time in notes by 9.5 percent against usual care, another showed no significant change, and physicians reported occasional clinically significant inaccuracies 17. In the London clinic where Wai was first built and used, and everywhere it has gone since, the doctor signs every note and nothing reaches a patient until a named person has approved it. Wai publishes its research in the open, including evaluation protocols, a documentation study across 12 clinics, work on deidentification and benchmarks against the standard of care.

What becomes possible
In 2024, an estimated 4.9 million children died before their fifth birthday, and the WHO and UNICEF report that most of those deaths could have been prevented with proven, low cost care 18. Behind most of them is a fever or a cough that someone saw and could not read in time. Put the judgment of an experienced pediatrician behind a community health worker in a remote district, so she recognizes the danger signs, and get the test result to someone who can act on it that afternoon, and a child who would have died goes to school. Many countries have already made the death of a small child a rare event. The medicine that would do the same everywhere is known.
In 2023, about 260,000 women died during or after pregnancy and childbirth, just over 90 percent of them in low and lower middle income countries, and the WHO says plainly that the ways to prevent or manage these complications are well known 19. A woman bleeds after the birth. Her blood pressure climbs in the last weeks of pregnancy. An infection starts quietly once she is home. Each has a treatment. What she lacks is a skilled person at the right moment, and bringing that skill to every birth is one of the most direct ways there is to give years back to a family.
Other illnesses take decades. Noncommunicable diseases caused at least 43 million deaths in 2021, 18 million of them before the age of 70, and 82 percent of those early deaths were in low and middle income countries 20. Heart disease, diabetes and many cancers respond to steady attention over years, a blood pressure measured and acted on, a dose adjusted, a follow up that actually happens. A record that never loses track of a person, read by a clinician who sees all of it, can move these numbers for whole countries.
More than a billion people live with a mental health condition 21. What helps them most is continuity, someone skilled who is there when they are needed and remembers them the next time.
Medicine is also changing shape as people live longer. By the mid 2030s, people aged 80 and over are expected to outnumber infants 1, and much of this century's medicine will be about keeping those added years healthy and independent.
What stands between these people and their care is mostly a shortage of skilled hands. The World Health Organization projects a shortfall of 11 million health workers by 2030 22, and the WHO and World Bank estimate that about 4.6 billion people still lack full access to essential health services 23. Training more clinicians matters, and it takes a generation. Sooner than that, we can change how far each clinician's skill travels, so that one experienced doctor's judgment supports hundreds of nurses and health workers, and each of them practices at the full standard of their training.
Our ambition is that preventable deaths become rare in every country within the working lives of the clinicians training today. We cannot promise the pace, so we will report our progress in public every year and let others judge it.
What the Foundation is for
No hospital or university can do this alone, because each works mostly with its own patients and its own questions. A network that spans countries can run one evaluation in many settings at once, and carry a finding from the hospital or clinic where it was made to every health system that can use it. The Wai Foundation exists to build that network among the world's researchers, clinicians and health leaders, and in The people who will build it I describe who we hope will join. We have not yet settled how the Foundation will be governed or which questions it will take on first, and we will decide both in the open.
Wai will keep publishing its research with the methods and limits stated, including results that do not favor us, and the Foundation will report each year on whom it has connected and what came of it.
The woman who came in breathless from her stairs should get that visit wherever she lives. So should the person waiting after her, and the one after that.
References
- United Nations Department of Economic and Social Affairs. UN projects world population to peak within this century, World Population Prospects 2024, July 11, 2024. https://www.un.org/sustainabledevelopment/blog/2024/07/press-release-wpp2024/
- Stanford Institute for Human-Centered Artificial Intelligence. AI Index 2025, State of AI in 10 Charts. https://hai.stanford.edu/news/ai-index-2025-state-of-ai-in-10-charts
- METR. Measuring AI Ability to Complete Long Tasks, March 19, 2025. https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/
- The Nobel Prize. Press release, The Nobel Prize in Chemistry 2024. https://www.nobelprize.org/prizes/chemistry/2024/press-release/
- American Medical Association. 2 in 3 physicians are using health AI, up 78% from 2023. https://www.ama-assn.org/practice-management/digital/2-3-physicians-are-using-health-ai-78-2023
- American Medical Association. More than 80% of physicians use AI professionally, AMA survey, March 2026. https://www.ama-assn.org/practice-management/digital-health/more-80-physicians-use-ai-professionally-ama-survey
- Stanford Institute for Human-Centered Artificial Intelligence. Inside the AI Index, 12 Takeaways from the 2026 Report. https://hai.stanford.edu/news/inside-the-ai-index-12-takeaways-from-the-2026-report
- David, P. A. The Dynamo and the Computer, An Historical Perspective on the Modern Productivity Paradox. American Economic Review 80(2), 1990, 355 to 361. https://econpapers.repec.org/RePEc:ags:uwarer:268373
- Brynjolfsson, E., Rock, D., Syverson, C. The Productivity J-Curve, How Intangibles Complement General Purpose Technologies. American Economic Journal, Macroeconomics 13(1), 2021. https://www.aeaweb.org/doi/10.1257/mac.20180386
- Morris, Z. S., Wooding, S., Grant, J. The answer is 17 years, what is the question, understanding time lags in translational research. Journal of the Royal Society of Medicine 104(12), 2011. https://pmc.ncbi.nlm.nih.gov/articles/PMC3241518/
- NHS England. Clinical Risk Management, its Application in the Manufacture of Health IT Systems (DCB0129). https://standards.nhs.uk/published-standards/clinical-risk-management-its-application-in-the-manufacture-of-health-it-systems
- NHS England. Guidance on the use of AI-enabled ambient scribing products in health and care settings. https://www.england.nhs.uk/long-read/guidance-on-the-use-of-ai-enabled-ambient-scribing-products-in-health-and-care-settings
- Assistant Secretary for Technology Policy, Office of the National Coordinator for Health IT. National Trends in Hospital and Physician Adoption of Electronic Health Records. https://www.healthit.gov/data/quickstats/national-trends-hospital-and-physician-adoption-electronic-health-records
- Arndt, B. G., et al. Tethered to the EHR, Primary Care Physician Workload Assessment Using EHR Event Log Data and Time and Motion Observations. Annals of Family Medicine 15(5), 2017, 419 to 426. https://pubmed.ncbi.nlm.nih.gov/28893811/
- Sinsky, C., et al. Allocation of Physician Time in Ambulatory Practice, A Time and Motion Study in 4 Specialties. Annals of Internal Medicine 165(11), 2016, 753 to 760. https://pubmed.ncbi.nlm.nih.gov/27595430/
- Kaiser Permanente Division of Research. AI-assisted notetaking gains steady support from Kaiser Permanente physicians, reporting Tierney, A., et al., NEJM Catalyst, 2025. https://divisionofresearch.kaiserpermanente.org/ai-assisted-notetaking-gains-steady-support-from-kaiser-permanente-physicians/
- Lukac, P. J., et al. Ambient AI Scribes in Clinical Practice, A Randomized Trial. NEJM AI 2(12), 2025. https://pubmed.ncbi.nlm.nih.gov/41497288/
- World Health Organization. Progress in reducing child deaths slows as 4.9 million children die before age five, March 18, 2026. https://www.who.int/news/item/18-03-2026-progress-in-reducing-child-deaths-slows-as-4.9-million-children-die-before-age-five
- World Health Organization. Maternal mortality, fact sheet. https://www.who.int/news-room/fact-sheets/detail/maternal-mortality
- World Health Organization. Noncommunicable diseases, fact sheet, September 25, 2025. https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases
- World Health Organization. Over a billion people living with mental health conditions, services require urgent scale-up, September 2, 2025. https://www.who.int/news/item/02-09-2025-over-a-billion-people-living-with-mental-health-conditions-services-require-urgent-scale-up
- World Health Organization. Health workforce. https://www.who.int/health-topics/health-workforce
- World Health Organization. Most countries make progress towards universal health coverage, but major challenges remain, WHO and World Bank report finds, December 6, 2025. https://www.who.int/news/item/06-12-2025-most-countries-make-progress-towards-universal-health-coverage-but-major-challenges-remain-who-world-bank-report-finds