Independence, AI, and Indigenous Sovereignty - Michael Running Wolf
AI can help preserve Indigenous languages, but only if we fight for sovereignty over data and culture
Hey you,
If you missed our episode with robot ethicist Kate Darling—complete with funerals for robot dogs and a truly cursed deepfake of yours truly—go back and watch it before your vacuum cleaner asks why you’re ignoring it. Now, on to our latest drop.
It’s July 4th week, and while many celebrate independence, I’m thinking about sovereignty and interdependence.
On sovereignty, I don’t just mean the right of a nation to assert itself and for its people to do what they want, but I mean data sovereignty and cultural sovereignty: the right of people to exist, speak, and thrive without being extracted, erased, or assimilated by someone else’s empire—whether that empire is a nation or a tech company.
It’s become an increasingly personal and urgent message as I’ve learned more about how the United States came to be. Not just the genocide of indigenous peoples–we know that story of harm–but the generosity of those people in offering the European settlers and colonists a model of government of which we are so proud and which could not have existed without native peoples: democracy.
The Haudenosaunee Confederacy (popularly known as Iroquois) represented Six Nations living under the Great Law Of Peace. This confederacy practiced democracy. They had hundreds of years of relations with various European colonies and literally advised the rebels who would become Founding Fathers of the USA. For the past year and a half I’ve been learning directly from several Haudenosaunee, Taino, and Lakota elders about the relationship between indigenous nations and the USA, and it’s clear that the independence of the USA would not have been possible without interdependence with the people already here. For a deep taste of the waters I’ve been wading in, check out the film project The Eternal Song. It’s a beautiful telling of the story of colonialism, trauma, and healing, and I was part of the online festival in conversation with Dr. Jose Barreiro, editor of Indian Roots of American Democracy.
These first people, whom we’re often taught live only in history books, are still here. Despite hundreds of years of efforts to undermine their sovereignty and erase them physically and culturally. That belated, humbling, and powerful education makes me all the more excited to share our episode with Michael Running Wolf. Every 14 days, the world loses another language. It’s an emergency hiding in plain sight, and Michael is tackling it head-on with VR, small language models, and AI tools to save Indigenous languages on the brink of extinction.
Our conversation was emotional, educational, comedic, and inspiring. Michael, who grew up on a very low-tech reservation, is using new high tech tools to unlock ancient wisdom we all need to heal a hurting world. He’s very aware of the risks of what he’s doing and demands that the tech industry and all of us treat data like land.
“What if you treat language like it’s land? It’s valuable. It’s something you own. So it’s data. Data is land.”
Watch the full episode here:
You can also listen on your favorite podcast platform.
Thanks for reading Life With Machines! Subscribe for free to receive new posts. Or subscribe with money to support the show and unlock bonus content and deeper analysis with a paid subscription.
Baratunde’s Take
Three things I haven’t let go of after this episode:
(1) Language extinction is as devastating as species extinction, and indigenous peoples are key to both.
You know how climate folks warn about losing a species every few days? We’re losing a language every two weeks. And when a language dies, it takes with it an entire ecosystem of knowledge, cultural memory, and a unique way of seeing the world. They also do more than describe. Languages encode. They hold. And they create the world. The words we use affect our reality and our values. The stories we tell create our experiences and shape our trajectory into the future.
But here’s the part that blew my mind: Indigenous languages aren’t just cultural artifacts; they are technological marvels. Many have built-in error correction—like TCP/IP but for human speech—making them resilient and precise.
“The very structure of these languages allows, if you only hear part of the word, you can reconstruct the rest.”
Michael explained that they are polysynthetic. The languages themselves communicate much more specific information than English (e.g. gender, family position, geographic context, and more). You can lose some pieces while maintaining the meaning. That’s a robust data infrastructure and gives hope to many of us whose histories have been erased that we might restore them. It also echoes a lesson I learned making America Outdoors with PBS. I spent time with indigenous communities who had brought back water sources and even species thought to be extinct.
While it may not be technically accurate to say that 80 percent of the world’s biodiversity exists on the 25 percent of the land stewarded by indigenous people, there is far less natural degradation on indigenous land than not. Many of our climate solutions lay in indigenous traditional ecological knowledge. The same may be true of our tech-based solutions, and language is one of the keys.
(2) AI can save languages—or erase them faster.
Michael is using AI to build tools that help kids and families learn their languages again. But there’s a catch. AI companies have realized that including Indigenous language data makes their models better. So they’re scraping that data without consent, turning endangered cultural treasures into fuel for private profit. The recent news about Meta’s ridiculous compensation packages for top AI talent drives the point home that we are in a new colonial race to dominate the future much like the European expeditions of old.
“If we do not participate in AI, it’s going to be another colonial tool.”
This is how extraction happens in the digital age: not by stealing land, but by harvesting voices. It takes the idea of “culture vultures” to a new level. And for people who have already suffered so much cultural erasure due to capitalism-fueled greed and Christian and white supremacy, this is all too familiar and must not repeat itself.
(3) We need a tribal data confederacy.
Michael’s vision is to build tribally-owned compute infrastructure so Indigenous communities can control their data, participate in the AI economy, and maintain sovereignty in a world that increasingly equates “participation” with “exploitation.”
“Data is land, and we should not give it away for free.”
This idea echoes efforts like Mozilla Common Voice and recent EU proposals on data sovereignty, but it goes further. It’s about building new systems that ensure AI isn’t just “diverse” but is accountable to the communities it touches.
It is a full circle return to the communities of interdependence that existed before the colonists landed in the “New World,” and its inspiring to meet someone from modern times using future tools to restore ancient and needed ways of being.
What Comes Next?
This conversation isn’t just about Indigenous languages. It’s about how we build a future where tech serves people, rather than the other way around. It’s about moving from “how can AI save culture?” to “how can culture save AI?” And it’s about how we, as people living in a world that’s getting more digital by the second, choose to demand accountability, consent, and sovereignty in every layer of that world.
If we want independence worth celebrating, it’s going to require more than fireworks and flag-waving. It’s going to require collective action, new systems, and a willingness to build differently.
Team Recommendations
The First Languages AI Reality (FLAIR) initiative Michael is part of and which could use your support or of that of any philanthropists and good companies you know
Guardian piece on the endangered language crisis
Mozilla Common Voice is a new open source community language model effort I learned about just two weeks ago when I opened and closed the Shared Futures forum produced by Aspen Digital and the Patrick J. McGovern Foundation.
Thanks for reading Life With Machines. Next edition will focus on the humans behind the machines. Lots of AI is just people!
Peace,
Baratunde