Dehcho AI Project

AI Built on Indigenous Land

The Dehcho region stretches across the Mackenzie Valley from the Nahanni Range to the Liard River — the traditional territory of the Dehcho First Nations. It is one of the most ecologically significant and culturally vibrant regions in Canada, and one of the most underserved in digital infrastructure. The Dehcho AI Project exists to build artificial intelligence capacity that originates in the Dehcho, serves the Dehcho, and is governed by the people of the Dehcho.

Problems That Require Local AI

The Dehcho faces challenges that demand local solutions: wildfire prediction across vast boreal landscapes, water quality monitoring for the Mackenzie River system, wildlife population management amid climate change, and language preservation for Dene Zhatie and South Slavey. These are not problems solvable by a model trained in San Francisco. They require data from the land, knowledge from the Elders, and compute infrastructure that stays in the territory.

OCAP and Sovereign Compute

The principles of OCAP — Ownership, Control, Access, and Possession — were developed for research data governance. Their application to AI is even more critical. AI models encode the values and biases of whoever builds them. When AI systems affecting Indigenous communities are built outside those communities, OCAP is violated at every layer. The Dehcho AI Project is designed to be OCAP-compliant: Indigenous ownership of infrastructure, Indigenous control of data, Indigenous access to models, Indigenous possession of hardware.