An independent institute · Founded 2026
Deployment is not transformation.
The Institute for AI Translation in Health Systems stewards a new discipline: a durable process for building the capacity, in institutions, communities, and individuals, to absorb each advance and apply it in their own context, in coherence with their values and their mission, so that everyone can benefit from the abundance this technology makes possible. The Institute holds the methodology, publishes the evidence, and funds the work of translation on the ground.
The systemic chasm
95%
of enterprise generative AI pilots deliver no measurable return
MIT NANDA · 2025
42%
of companies abandoned most of their AI initiatives in 2025, up from 17%
S&P Global
80%+
of AI projects fail, roughly twice the rate of comparable non-AI projects
RAND
52%
of workers are worried about AI in their workplace
Pew Research · 2025
The failures are not technology stories. They are stories about people who were never brought along, and the successes are stories about people who were. The full evidence library →
The discipline
AI Translation is the discipline of making explicit what is usually left unexamined: the beliefs an institution actually holds, the assumptions beneath them, how those beliefs shape thinking, and how thinking moves into action. Its first principle is that there can be no unity of action without unity of vision and thought. Its instrument is a set of four questions, asked of every project, before and during the work.
The discipline rests on years of comprehensive study and experience of how technology is designed, disseminated, applied, resisted, and, most instructively, of how it fails. The question it answers is not how to deploy. It is how to create a durable process: one in which the right use cases are identified by the people who must live with them, and trust emerges from a pattern of coherence between what is said and what is done.
The innovation process is creative and yet structured; so is the translation of innovation into complex systems. Capacity is cultivated in a learnable sequence, across institutions, communities, and individuals together, and the learning is shared as it accumulates. Progress is governed accordingly: translations advance by developmental milestones, not time-based ones. A phase is complete not when a date arrives but when the qualities and capabilities it exists to cultivate have demonstrably emerged, and the discipline's craft is creating the conditions in which they do.
Does this enhance the institution's purpose, its vision, and who it understands itself to be?
Can this be built and operated within the real constraints of this place, now?
Will this sustain itself, economically and institutionally, after the initial energy has passed?
Can the knowledge, judgment, ownership, and capability required to sustain it become rooted locally over time?
Theory of change
The three aims · the centerpiece of the theory
I
To enable universal access to, and participation in, the benefits of frontier AI, closing the digital divide from both sides: communities able to use the technology, and ventures able to serve them.
II
To build the capacity of health systems and communities to apply AI in coherence with their values, so that technology serves human flourishing on terms each community defines.
III
To remove the financial and technical barriers that keep builders from the problems that matter most, so that the worth of a problem, not the size of its market, determines whether it is solved.
Venture economics rightly seek scale; that is how frontier technology comes to exist at all. But the same economics pass over district hospitals, community clinics, and local institutions, markets too small for venture capital and too important to ignore. Different social missions require different economic architectures. The Institute's answer is to fund and govern translation through a structure designed for permanence rather than exit: services co-created with local builders, operated by local teams, and owned by the communities they serve, with surplus recirculating to seed the next region. The model is non-extractive by construction: no equity taken, no value siphoned away, capability compounding where it is built.
The Institute's delivery instrument is doáb, a development cooperative: a network of country AI Translation centers, each cultivating an ecosystem in which local builders develop and launch ventures that solve the problems around them, ventures that stay small, reach far, and join together when the work is larger than any one of them. The Institute takes no equity in what is built, because the outcome being measured is capacity: every community able to generate knowledge, to apply it, and to take part fully in shaping its own future.
How institutions engage
Each is described in full on the Engage page.
A bounded decision process that turns AI debate into a sequenced, stakeholder-aligned portfolio. Nothing moves to delivery until priorities are genuinely shared.
Hands-on delivery in short cycles with the institution's own people, ending with services in production and a team whose judgment is its own.
A multi-year program that builds a country's or region's own delivery capability, with transfer criteria agreed before implementation begins.
The front door of every new geography is a Health Systems Innovation Hackathon: methodology and judging carried with academic partners, local mobilization carried by the cooperative, and every promising team flowing into accompaniment rather than a demo day and a goodbye.
The evidence program
Every engagement is documented truthfully, including what does not work; negative results are recorded and published with the same care as success. The first countrywide demonstration is under way and is being stewarded quietly to protect the focus of local partners; complete, open-source case studies are published as milestones are reached. The Institute's claims are bounded by its Academic Council: nothing is presented as evidence that has not met its standard.
A shared vocabulary
The discipline of fitting frontier AI to the conditions, values, and aspirations of a community: making explicit the beliefs, assumptions, and thinking beneath action, so that capability becomes capacity.
Not hospitals alone: the dynamic whole of people, institutions, and resources whose primary purpose is to protect and improve health. A complex, adaptive system whose functions, governance, financing, resource generation, and service delivery, are judged by better health and its fair distribution, protection from financial hardship, responsiveness to people, and efficiency.
The structures through which a community acts beyond the individual: ministries and municipalities, hospitals and universities, agencies and associations. Institutions are carriers of continuity; their capacity determines what a society can sustain.
A body of people bound by place and shared concern, possessing its own knowledge, values, and aspirations. In this discipline the community is never the object of a program; it is the protagonist of its own development.
The cultivation of a people's power, in individuals, communities, and institutions together, to shape their own future: skills and abilities, and equally the qualities and attitudes that govern their use, with opportunities for practice so that cycles of action, reflection, and feedback turn experience into knowledge held locally.
Walking alongside rather than doing for: strengthening judgment and capability until presence is no longer needed. The measure of accompaniment is what remains after departure.
A structure in which purpose-built ventures stay small, reach far, share infrastructure and learning, and join together when the work is larger than any one of them, with surplus recirculating to seed capacity where it does not yet exist.
Progress measured by the demonstrable emergence of qualities and capabilities rather than by dates. A phase is complete when what it exists to cultivate has appeared.
Governance
The Institute is a nonprofit corporation governed by an independent board. Its research agenda and methodological standards are held by the doáb Academic Council, whose Founding Chair carries authority over what the Institute may claim as evidence. Financial boundaries between the Institute, its academic partners, and the cooperative are published and kept clean by design.
Convenes scholars and practitioners across health systems, technology, and development. Sets the research agenda, reviews the evidence program, and safeguards the discipline's standards as the work scales across geographies. The Council in full →
Partner with the Institute
Philanthropic grants, recoverable grants, and sponsored builder seats, deployed under a published discipline, with the Moonshadow Fund as the Institute's philanthropic vessel.
Frontier tools, compute, and expertise extended to the places they do not yet reach, so builders everywhere can apply them and contribute learning back.
Introductions, convenings, and platforms. Some of the most valuable contributions to this work are made without writing a check.