Faced with a tech industry built on extraction, Indigenous technologists are exploring ways to build tools while protecting cultural and data sovereignty.
At the Indigenous AI Gathering at Montréal AI institute Mila this week, Indigenous scholars, tech workers, and AI researchers delved into the tension between generative AI’s significant use of data and the ways it can be harnessed to build solutions to socioeconomic problems. The two-day conference featured presentations from the Indigenous Pathfinders in AI program—a seven-week educational program at Mila that gives Indigenous folks intensive machine learning training and focuses on Indigenous-centred approaches to the tech.
“The essential problem that we come across in AI is: where are you getting the data, and how are you making sure that’s community-informed?”
Program participants pitched the AI platforms they had built, and across the presentations, data sovereignty was built into the platforms: nothing leaving the network, meant to run locally within the communities for which they’re built. Pitches included Landlens, which analyzes accrued impact of land developments through public data; SAI Cheese, a preventative dental care app where health data is stored locally; and Portage, a K-12 education platform for teachers trained on cultural knowledge.
“The essential problem that we come across in AI is: where are you getting the data, and how are you making sure that’s community-informed and consented?” said Ali Lang, a senior accessibility analyst (Indigenous) at the CBC who co-created Portage during the Pathfinders in AI program, in an interview with BetaKit. “There’s a data governance onboarding layer that’s our foundation.”
One specific example was the extraction of personal data used in AI for healthcare. Arizona State University geneticist Krystal Tsosie, of the Diné/Navajo Nation, outlined the major problem in datasets for healthtech research: for-profit companies will utilize data sourced from Indigenous communities, but then often make final healthcare innovation inaccessible or too expensive for the community to access.
“[Data] determines what the system sees, what it optimizes for, who benefits, and who is harmed,” Tsosie said. “Our community’s data is already on this pathway, and the question is whether or not our community governs what it becomes.”
Tsosie argued for research models that actually respect health-data sovereignty—which includes having samples and data housed on ancestral lands; research questions being set by community members; and putting community benefit as the design goal rather than an afterthought.
RELATED: ITC says Canada’s largest Indigenous tech conference aims to break down silo
In Canada, there is a long history of the federal government misusing data to surveil, exploit, and even erase information about Indigenous communities—which has led to standards such as the First Nations Principles of OCAP to govern how data can be used. At the Mila event, AI and digital innovation minister Evan Solomon acknowledged the “scar tissue” between the federal government and Indigenous communities when it comes to data use. Solomon was gifted the sacred herb of braided sweetgrass by University of Calgary doctoral candidate and Mila fellow Mick Elliott, who said it came with a “great weight and responsibility” toward Indigenous people.
Michael Running Wolf, co-founder of Mila’s First Languages AI Reality (FLAIR) program, acknowledged that building in AI might seem unattractive, given how large tech companies have used the tech for exploitative purposes. But he told the audience that the goal of having more Indigenous representation in AI “isn’t to contort ourselves, so that we break ourselves.”
“It’s that we demonstrate the flexibility of our culture and the behaviour as representatives of our culture,” he said.
All images courtesy Madison McLauchlan for BetaKit.