Nicole Horseherder has seen the impacts of unsustainable development on Indigenous communities. A Navajo environmental activist and co-founder of Tó Nizhóní Ání (Sacred Water Speaks), a Diné-led nonprofit organization based in Arizona in the US, she has spent years protecting water that sustains communities from industrial use.
She sees parallels with today’s artificial intelligence development, she said.
As technology is advancing at an unprecedented rate, a growing body of research is looking at Indigenous knowledge systems for guidance on ethical frameworks for AI. But for someone like Horseherder, Indigenous knowledge is not data to be harvested, she said.
“It is built on thousands of years of real-time human observations on the changes in landscapes, the weather and the seasons, the directions of the moon, the sun and everything around us,” she said. Within the Navajo community, people living in different landscapes including the high-deserts, river valleys and dry to arid places have their own local knowledge systems.
A recent study published in AI and Ethics journal examines how Indigenous ecological knowledge could reshape AI frameworks through an analysis of Navajo and Māori concepts. The paper drew on Māori value of Kaitiakitanga, or guardianship, and Navajo philosophy of Hózhó, meaning balance and harmony.
The study’s authors said that traditional ecological knowledge embodies collective responsibility and could provide an ethical basis for questioning whether the scale of a proposed AI model is justifiable given its environmental cost, prioritizing ecological integrity over unbounded technological expansion.
According to the proposed theory, incorporating traditional ecological knowledge has another benefit: AI models can accurately reflect the complex relationships between species and their environment, animal behavior and habitat use to avoid biases from relying solely on Western scientific data.
Jude Kong, an assistant professor at the University of Toronto, who studies community-oriented AI and public health, put it plainly: “You need to learn from local communities what their problems are. Otherwise, you are moving into this colonial way of saying ‘This is your problem and this is your solution.’ That has never worked,” he said in a virtual meeting with Mongabay.
“
AI is a great technology and has an astounding learning rate. Maybe it can provide information on Indigenous communities, but it will never be an extension of our existence.
Nicole Horseherder, co-founder, Tó Nizhóní Ání
According to Kong, who was not an author of the study, AI frameworks often fail to gain the trust and acceptance of local communities when designed and deployed without consulting them.
Including Indigenous knowledge
Research suggests that AI could become a powerful ally in conservation when developed in partnership with Indigenous communities. Already, the authors said, AI is being integrated into several ecological monitoring programs in collaboration with local communities around the world including parts of the Amazon rainforest. AI tools are used in identifying the causes of deforestation in Congo Basin and Indonesia, while also tracking the spread of illegal gold mining in Amazon.
In addition to its use in conservation, the authors said that AI development can be guided by Indigenous values such as Kaitiakitanga and Hózhó, which emphasizes stewardship, reciprocity and living in balance with nature rather than treating it merely as a resource for human exploitation.
However, the authors said governance mechanisms proposed in the study remain theoretical until they are validated, critiqued or refined by the communities themselves.
In another similar study in 2024, Sebastián Lehuedé, a digital humanities researcher and lecturer in ethics, AI and society at King’s College London, UK, looked at communities’ resistance against the infrastructure that underpins AI through a proposed data center by Google in Chile; data centers can often add additional stress on water resources. He observed that Indigenous peoples and local communities see resources like water not just from the point of ownership but also as an agent present in their everyday life and as an enabler of a way of being.
Lehuedé said that ethics in AI development goes beyond a yes or no approach to the technology and is rather rooted in whether its resourcing disrupts the relations, including between people and nature, that sustain the planet. Indigenous values in the Lickan Antay communities, for example, oppose the transformation of natural elements like water, air and earth into taken-for-granted AI infrastructure, while encouraging care and obligations toward land and communities.
The proposed Google data center ultimately did not see the light of the day in Chile due to widespread protests from the locals.
In both studies, the authors highlighted that AI tools, infrastructure or value chains that exclude the values of Indigenous and local communities often face backlash, highlighting the need for reframing the ethical considerations around AI. The studies mandate Indigenous participation in AI policies and establishing detailed community impact assessments before deploying AI projects in Indigenous territories.
Limits in AI
Karaitiana Taiuru, an independent expert of Indigenous data sovereignty who also heads Taiuru and Associates Ltd, a New Zealand-based consultancy that provides services in AI projects in the Māori context, had some reservations about the study published in AI and Ethics.
He said that engineers and developers from non-Indigenous communities writing ethical frameworks on Indigenous data in peer-reviewed journals is a form of digital colonialism, much like historical colonial extraction.
“Treating Kaitiakitanga and Hózhó as equivalent units in a comparative qualitative analysis is itself a methodological expression of colonialism,” Taiuru said, adding that the study reduced culturally grounded, but distinct Indigenous principles into a generic “Indigenous knowledge systems” category. “The specific, irreducible, culturally grounded content of both systems is precisely what gets lost in this process,” he said.
Mongabay reached out to the authors of the study for a comment, but did not receive a response by the time of publication.
Taiuru said that despite these challenges, Indigenous communities cannot afford to remain absent from AI governance models that are influencing environmental monitoring systems, healthcare and education. According to him, the question is no longer about Indigenous representation in AI but whether Indigenous people will have meaningful control over how AI systems are designed and deployed.
“My fear is that if we don’t speak out, we will be left behind,” he said.
Taiuru described himself as an Indigenous data sovereignty optimist. “We need to decide what data can be used and what data should never be used in AI,” he said. “A gun is dangerous only as long as a human pulls the trigger. It’s the same with AI.”
But for some like Horseherder, such technological projects miss something more fundamental. She said she simply cannot see how Indigenous knowledge, shaped over thousands of years, could be translated into a technology that is still evolving.
“AI is a great technology and has an astounding learning rate. Maybe it can provide information on Indigenous communities, but it will never be an extension of our existence,” she said.
This story was published with permission from Mongabay.com.
