Indigenous-Led AI: How Indigenous Knowledge Systems Could Push AI to be More Inclusive

New Frontiers in Research Fund | Published:

Talking Circle at the Indigenous Protocol and Artificial Intelligence Workshops
Photo: ʻĀina Paika | Courtesy of the Initiative for Indigenous Futures, 2019

As the power of artificial intelligence begins to transform the world, an Indigenous-led research team is mobilizing to ensure AI is inclusive of Indigenous Knowledge systems.

“This is arguably the first major technological revolution in the West where Indigenous Peoples have the capacity to fully participate in its shaping,” says Jason Edward Lewis, who is a professor of computation arts at Concordia University as well as the University Research Chair in Computational Media and the Indigenous Future Imaginary. “We have an opportunity here to affect the trajectory.”

Lewis is co-leading the Abundant Intelligences project, which was recently awarded a New Frontiers in Research Fund Transformation grant worth more than $22 million. It is one of six large-scale, interdisciplinary research projects that will address major challenges over the next six years with funding support from the Government of Canada. The project’s co-principal investigator is linguist Hēmi Whaanga of Te Pūtahi-a-Toi, the School of Māori Knowledge, at Massey University in New Zealand.

We don't have an AI ethics problem, we have an AI epistemology problem.

In total, the project’s interdisciplinary team of experts includes 37 co-investigators and collaborators who come from eight universities and 12 Indigenous community-based organizations in Canada, the United States and New Zealand. Most of the team members are Indigenous. They are motivated to expand the definition of intelligence by collaborating with Indigenous communities to integrate their knowledge systems with the AI research and development ecosystem.

“This is a culmination of many years of conversations,” says team member Suzanne Kite, an artist in residence and visiting scholar at Bard College in New York. Kite studied under Lewis as a PhD student at Concordia, where her academic work imagined new ethical AI protocols that examine past, present and future Lakȟóta philosophies.

Now, with commercial applications of AI taking off, time is of the essence. Lewis sees this current phase of AI as the most consequential yet. He began his professional career in Silicon Valley, an experience that led him to start asking big questions about how worldview affects what kind of technologies we discover and develop. While the problems surrounding bias in AI are well-documented, he says the root cause is much deeper and not getting the same level of attention.

“We don't have an AI ethics problem, we have an AI epistemology problem,” Lewis says.

Raymond Thomas

Participants in the Indigenous Protocol and Artificial Intelligence Workshops (Jason Edward Lewis third row, middle, wearing a navy polo shirt)
Photo: ʻĀina Paika | Courtesy of the Initiative for Indigenous Futures, 2019

Lewis, who is Hawaiian and Samoan, shares an example: a normative western approach that is favoured by current AI research is the assumption that the user is an individual, and the individual prioritizes their own well-being. He says this creates a blindness to vital aspects of human existence—such as trust, care and community—that are fundamental to how intelligence actually operates.

“It creates huge gaps between the diverse forms of human and nonhuman intelligence and what the AI industrial-academic complex is trying to replicate,” Lewis says. “Collapsing intelligence down to a rational, goal-seeking, self-serving agent—in all of our cultures, that kind of person would be seen as selfish, foolish, not a good community member, and not intelligent.”

Indigenous epistemologies (or theories of knowledge) provide frameworks for understanding how technology can be developed in ways that integrate it into existing ways of life, support the flourishing of future generations, and are optimized for abundance rather than scarcity.

For example, when Whaanga discusses what it will mean to articulate, shape and design AI through a Māori lens, he talks about putting family first.

“We aim to mobilize AI technology to explore the centrality of hapū (kinship group) and connection in creative, language, cultural and well-being contexts,” Whaanga says.

The idea that the field needs to expand its definition of intelligence is attracting attention, including at Mila, the AI research institute, based in Montréal, Quebec, which is a partner on the project.

“Many colleagues from the AI field are beginning—better late than never—to realize that, in addition to developing a more ethical and equitable AI framework, building AI systems that learn from more diverse perspectives and values will result in better and more robust AI,” says Karim Jerbi, who is a Mila associate research member and Canada Research Chair in Computational Neuroscience and Cognitive Neuroimaging at the Université de Montréal.

“The goal is to broaden AI and bring a more holistic and inclusive approach to the field,” Jerbi says.

… building AI systems that learn from more diverse perspectives and values will result in better and more robust AI.

To achieve this, the team came up with the idea of “pods,” which are base camps for deeper engagement with different Indigenous communities. In Alberta, for example, Jackson Leween, Two Bears, a multimedia installation and performance artist and Canada Research Chair in Indigenous Arts Research and Technology at the University of Lethbridge, will lead a Blackfoot/Mohawk pod. In the US, Kite is forming a research pod of community members, scientists and students who will delve into Indigenous research creation and, for example, apply Lakȟóta design principles to AI wearable devices.

The Abundant Intelligences team is excited about the spin-off benefits of engaging Indigenous communities in AI’s development. It’s a way to develop engineering and design capacity within and across global Indigenous communities and map out an ethical and robust development path for Indigenous communities to create and use such systems.

While the primary impact will be to benefit Indigenous communities in North America, the Pacific and elsewhere in exerting sovereignty over their computational landscape, Lewis envisions that the project will also benefit society in general by better equipping us to adapt to both the challenges and benefits of AI.

“There are big wins to be had from this project, as we shape this new and incredibly important technology in a direction that’s better for Indigenous Peoples and better for mainstream society, too.”

Date modified: