About
Research that earns its place.
Hasna Foundation is a research foundation. We support work that connects AI to problems with public consequences: how care is delivered, how climate risk is understood, how children learn, and who gets to take part in science.
We are deliberately narrow. A short agenda, a small number of programs, and a preference for work that can be checked. The foundation exists to make careful research possible, not to sell a product or promote a technology.
Why it exists
Some of the most important AI research will not be funded by markets. Its benefits are widely shared, its results arrive slowly, and its value is difficult to capture in a product. That is not a failure of the research. It is a mismatch between the work and the way most funding is organized.
Hasna Foundation is built around a different set of incentives. The foundation starts from problems that affect large numbers of people, asks for results to be published, and stays with a question long enough to learn whether an approach is working, including when the answer is no.
The foundation is designed to start small: a focused agenda, a short list of programs, and partnerships with people who already understand the problems. A broad strategy announced before the work exists would help no one, so we would rather do a few things carefully and report on them honestly.
The foundation is meant to last longer than a single grant cycle. Problems in health, climate, and education do not resolve in a year, so we plan for repeated rounds of work, each one reporting what changed and what did not.
Values
What we hold ourselves to
- 01
Openness
Methods, code, and results are published unless there is a specific reason not to, such as participant privacy or a safety risk. When we withhold something, we say why.
- 02
Rigor
We prefer careful, reproducible work to an impressive demonstration. A claim should survive a second attempt by someone else.
- 03
Stewardship
Resources entrusted to the foundation are a public trust. We spend deliberately, avoid overhead for its own sake, and account for what we do.
- 04
Dignity
People affected by a problem are participants in the research, not an audience for it. Their constraints and their judgment shape the work.
Approach
- Start from the decision the work is meant to improve, not from a technique looking for an application.
- Fund teams that can stay with the problem long enough to learn something real.
- Publish failure as part of the record. It is often the most useful thing a project produces.
- Measure what changed outside the lab, not only what improved on a benchmark.
Structure
Governance
The foundation separates direction, review, and delivery so that no single role decides what counts as good work. The structure below describes how those responsibilities are divided.
Board
Sets the mission and annual budget, approves the research agenda, and is accountable for keeping the foundation focused on public benefit.
Research Council
Reviews proposals against the agenda and advises on technical direction. Members are researchers and practitioners from outside the foundation.
Program leads
Run grant rounds and fellowships, support grantees, and make sure results reach the public record.
Operations
Handles finance, administration, and compliance, and keeps the public record of grants and spending accurate.
- Grant decisions and the reasons for them are documented and shared with applicants.
- Conflicts of interest are disclosed, and anyone with a conflict is recused from the relevant decision.
- A yearly summary of grants and spending will be published once the first grant round closes.