Governance, ethics, and data rights
The technical questions around AI for the living world are inseparable from a set of older, harder questions: who owns what was observed, who gets to decide what it means, and who is accountable when a model is wrong. We treat these as research problems, not afterthoughts.
What we study
- Community data rights. How rangers, fishers, farmers, pastoralists, and Indigenous and local communities can contribute observations without surrendering ownership of them — and what licensing, attribution, and consent look like in practice.
- Auditability of findings. What it takes for a published map, indicator, or prediction to be checkable months or years later — by another lab, a regulator, or the community it concerns. Methods, datasets, model versions, and assumptions all need to be reachable from the result.
- Where AI should and should not decide. Honest scoping of which decisions are appropriate to automate, which need a person in the loop, and which should not involve a model at all. The default for high-stakes ecological judgements is human judgement, supported — not replaced — by AI.
- Failure modes and accountability. What happens when a model is wrong, when a dataset is biased, when an indicator misleads a decision-maker. Who notices, who fixes it, who is responsible.
Why this matters
Ecological evidence increasingly drives consequential decisions — about protected areas, fishery quotas, restoration funding, food security policy. The credibility of those decisions depends on the integrity of the evidence chain underneath them, and on the trust of the communities and institutions whose work it represents.
Without clear governance, even the best methods can produce conclusions that quietly harm the people they were meant to help. Without clear data rights, communities have little reason to participate. Without auditability, no one outside the original project can tell whether to believe what was published.
How we work
We approach governance the same way we approach ecology: with care, with humility, and in collaboration with the people most affected.
We work openly with researchers, legal scholars, and community representatives. We publish frameworks rather than impose them. We treat governance not as compliance but as part of the science.