Governance
How we decide things, how positions change, and who curates the work.
India has already built most of what a population-scale health system needs: digital identity, health records, public insurance, one of the largest immunisation programmes in the world, and close to a million community health workers. What it does not yet have is a layer that answers, for one person at a time, what should happen next.
IHMR, the India Health Maintenance Rail, is an open research project asking what it would take to build that layer - infrastructure that continuously helps maintain health, instead of a system that waits for illness and then reacts to it.
IHMR Labs is a collective of people interested in where AI, health, and population-scale impact meet: clinicians, engineers, researchers, and people who have worked inside the health system and know where it actually bends. We publish in the open, including the parts we have not worked out.
It is headed by Tejas Parthasarathi Sudarshan, who convenes the work and curates what gets published under the IHMR name. There is no institutional funding, sponsorship, or government affiliation behind any of it, and how decisions get made is written down.
The project is early, and most of what is here is still hypothesis rather than settled. We would rather say that plainly than imply a standing we have not earned. If you know something we have got wrong, that is the most useful thing you can bring us.
How we decide things, how positions change, and who curates the work.
Where AI is used, how we disclose it, and why a named human signs off on everything.
How comments are reviewed, why we review all of them, and why disagreeing with us is never a reason not to publish.
Where this is going, and how early we really are.
Our vocabulary, including the words we are still unsure about.
Worth reading. There are several things people reasonably assume that are not true.
What we collect, why, where it lives, and what you can ask us to do.
How we protect it, and what we deliberately do not hold.
The two we set, and the ones we only set if you agree.