AI Is Not Coming for Your Culture

Recently, in a discussion with a youth educator at one of the Native American Organizations we do business with, I was stunned when he made it clear that he does not want to promote AI at all for the youth of his NAO. To say the least, I was stunned. I couldn't relate to the fear he was putting out. I have always been eager to learn more about computer technologies, ever since it became more public friendly in the early 1970s. Every new machine that came along was a door, not a wall. So when a thoughtful educator — a man who clearly loves his young people and wants the best for them — told me he'd rather keep this particular door closed, it stopped me in my tracks. We agreed to have a follow-up discussion later. Until then, it's been mulling in my mind. And the more I've turned it over, the more it occurred to me that he isn't alone. He's speaking for a lot of people. He's speaking for parents, program directors, culture keepers, and elders who have watched technology roll into their communities before, and who have learned to ask a very reasonable question first: What is this going to cost us? That is not fear. That is stewardship. And it deserves a real answer, not a sales pitch. So I went and looked into it. Here is what I found — and here is the sentence that ended up organizing everything else:

The question is not whether Native communities should accept AI. The question is whether Native communities will decide for themselves how, where, when — and whether — AI is used.

Everything below serves that one sentence.

1. What AI Actually Does — and What It Doesn't Become

There's a debate running right now among serious researchers about how creative these systems really are. One line of work, called the Einstein Test, asks a genuinely interesting question: could a machine, given only the information available before Einstein's insight, have produced the insight itself? Researchers disagree about the answer, and I'm not going to pretend the argument is settled. Here's the good news: you don't need it settled. Because the argument that matters to a Native Nation isn't about what a machine might eventually be capable of. It's about what a machine is — and what it is not. Today's AI systems are extraordinarily good at finding and generating patterns from the information and examples available to them. They can draft, summarize, translate, organize, explain, and suggest, at a speed no staff of any size can match. But generating patterns is not the same thing as human judgment, cultural authority, lived experience, or responsibility. A system can produce a beautiful explanation of a ceremony without carrying what that ceremony means to the people who keep it. It can translate a word without holding everything that word carries in one particular community. It can summarize a history it has not lived, on behalf of a people it does not belong to, for purposes no one asked it about. It has access to information. It does not automatically possess the relationships, responsibilities, cultural authority, lived experience, or community consent that give that information its meaning. That's the distinction. And it's the one worth remembering:

AI is strong at patterns. Humans are strong at meaning. I don't offer that as a scientific law. I offer it as the organizing idea of this whole discussion — and as the reason a machine does not become an Elder, a culture keeper, a Tribal leader, a parent, a teacher, or a citizen simply because it can produce convincing words about those roles. Native communities have long understood that information and wisdom are two different things — that a fact you can look up and a truth you have to live are not the same animal. The rest of the world is arriving

at that distinction now, late, by way of research papers. It has been foundational in Native life a great deal longer. So when that youth educator worries about a machine moving in and taking the place of what his community carries — the honest answer is that the thing he is protecting is not the kind of thing a pattern system can take. It can be copied. It can be misused. It can be misrepresented. Those are real risks, and we'll get to them. But it cannot be replaced, because replacement would require becoming something the machine has no standing to be.

2. Where AI Can Actually Help a Native American Organization

Here's the other half, and it's the half that changes the calculation. AI does not have to replace Native Nations. It can be used to serve Native Nations — if Native Nations decide how it will be used and stay in charge of the decisions. That second clause is the whole ballgame. A tool serves whoever holds it and sets the rules for it. Think about what most NAO staff actually spend their days doing. Reports. Grant narratives. Meeting minutes. Program summaries. Policy drafts. Letters to funders. Newsletters that never get written because there's no time. Intake paperwork that eats the morning. Turning a forty-page federal requirement into something a program director can act on before Thursday. Every one of those is pattern work. Every one of those is what these systems are good at. Used deliberately, AI can help: Programs move faster — a first draft in two minutes instead of two days Staff spend less of the week on paperwork — summaries, forms, and reports stop swallowing the calendar Elders get information in plain language — larger type, read aloud, no jargon, at their own pace Youth learn real skills — as operators of the technology rather than just consumers of it Organizations build capability that stays home — the skill lives with your people, not on a contractor's invoice Small teams do the work of larger ones — expand services without expanding the budget Educators teach at any reading level — the same material rewritten for a 7th grader or a grant reviewer in seconds

Governments strengthen administration — the legal, health, and administrative capacity scholars are now documenting That last point isn't just my opinion. Adam Crepelle, writing in the Duke Law & Technology Review, argues that AI offers real possibilities for strengthening tribal courts, healthcare delivery, education, cultural preservation, economic development, and governmental administration — and he opens by noting that tribes have been largely neglected in AI policy discussions, despite being sovereigns. Read that again. Not targeted. Neglected. Left out of a conversation being held about them anyway. Look closely at the list above. Not one item on it asks a community to give anything up. Every item is about capacity — doing more of what you already want to do, with less friction between the intention and the result. AI is not inherently a threat to sovereignty. It becomes a tool for sovereignty the moment a Nation decides how it will be used.

3. Sovereignty Means More Than Keeping Data Private

This is the part I'd underline if I could only keep one section. Sovereignty is not a technical question about where a server sits. It is a decision about who gets to decide. Who decides whether AI is used at all? Who decides what information it may touch? Who decides what must never be put into it? Who approves the final answer before it goes out? Who owns the resulting work? Who is accountable when the system gets something wrong? Those are governance questions, not technology questions. And Native Nations are entirely capable of answering them for themselves — the same way they answer every other governance question. Here's what makes this urgent rather than theoretical. As of September 2025, according to the University of Oklahoma's Native Nations Center, not one federally recognized tribe had enacted AI risk regulation. Not one, out of 574. That is not a criticism of anyone. It's a description of an open field.

The European Union passed a comprehensive AI law in 2024. Colorado and Texas have adopted state AI governance frameworks. The United States has no comprehensive federal AI statute. And in that gap, the question of how AI touches Tribal information, Tribal language, and Tribal culture is currently being answered — by default, by vendors, by state agencies, and by whoever happens to be typing. Every day that stays true, somebody else's defaults become your policy.

Capability is not permission Here is the idea I'd most like to leave with a Tribal administrator:

If information belongs to a community, the fact that a machine can technically process it does not mean the machine should be allowed to process it. Capability is not permission. That distinction — between what a technology can do and what a community consents to — may be the single most important AI governance question facing Indian Country right now. It is also the concern the specialists keep raising. The University of Oklahoma's work specifically flags the unauthorized use of Tribal cultural and traditional information for AI training, including situations where government employees simply paste Tribal information into a public AI tool because it was convenient that morning. Not malice. Convenience. Which is exactly why it needs a written rule rather than good intentions. The same principle is already formalized in the CARE Principles for Indigenous Data Governance — Collective benefit, Authority to control, Responsibility, Ethics — which exist precisely because "the data was available" was never the same thing as "the community agreed."

4. The Basics — and the One Question That Comes First

The good news is that the practical side is not complicated. You don't need to be a programmer. You don't need a computer science degree. You don't need to understand what happens under the hood, any more than you need to understand combustion to drive to work. Five habits carry an ordinary person a long way: 1

Ask clear questions. Vague in, vague out.

2. Use plain language. There's no secret vocabulary. Talk to it like a person.

3. Check the answers. It can sound completely certain and be completely wrong.

4. Know where your information goes — who can access it, whether it's retained, and whether it's used to train the system.

5. Use AI as a helper, not a decider. It drafts, suggests, and summarizes. People decide.

Those five habits won't solve every AI problem. But they give an ordinary person a remarkably good place to start. And for a Native American Organization, there's a sixth that comes before all of them:

6. Know what should never go into an AI system at all. Sensitive Tribal, cultural, ceremonial, personnel, health, legal, enrollment, or otherwise protected information should never be entered into an AI system simply because it is convenient. That isn't a technical rule. It's a governance rule, and it should be written down, approved, and taught — before anyone is handed a login.

Before we use AI: eight questions worth asking Any NAO can put these on one page and start using them Monday morning: 1

Who owns this information?

2. Is it already public?

3. Does it contain confidential, ceremonial, or culturally sensitive material?

4. Has the appropriate person or authority approved its use?

5. Where does the information physically go once it's entered?

6. Will it be retained, or used to train the system?

7. Who can access the output?

8. Who reviews the result before it is used or released?

Eight questions. No technical background required to ask a single one of them. And if the answer to any of them is "I don't know," that is itself a useful answer — it tells you exactly where your policy needs to start.

5. Awareness Matters — But It Doesn't Have to Mean Retreat

Let me be straight with you, because you'd see through anything else.

AI has potholes. Every technology does. It can state something wrong with complete confidence. It can produce a citation that doesn't exist. It can carry a bias in and hand it back dressed up as an answer. It can misrepresent a culture it has only ever encountered as text. More precisely, and more usefully than the sweeping version: AI does not independently verify reality just because it produced an answer. AI can sound certain when it is wrong — it has no reliable way to signal the edge of its own knowledge. AI does not automatically give you trustworthy provenance for the claims it makes. AI-generated novelty still requires human evaluation, judgment, and responsibility. Those are real limitations. They are also the exact reasons the sixth rule above exists, and the reason "helper, not decider" is the anchor of the whole approach. But notice what none of them says. None of them says the technology is beyond governing. AI is not magic. It is technology — and technology can be governed. The question is not whether AI will have consequences. It will. The question is who gets to decide what those consequences are. And on that front, something encouraging is happening right now. This is no longer a conversation Native Nations are absent from: The National Congress of American Indians and Arizona State's American Indian Policy Institute have jointly launched a Center for Tribal Digital Sovereignty. The University of Oklahoma's Native Nations Center is publishing an ongoing Sovereign Snapshot series on Tribal AI governance and AI in a Tribal context. Brookings and AIPI have published directly on defining digital sovereignty for Tribal Nations in the AI age. Legal scholarship in the Duke Law & Technology Review is examining how AI can strengthen — not merely threaten — Tribal sovereignty. This isn't a fringe worry anymore. It's an organized field with institutions in it. And there is still plenty of room in the room. Which brings me to the sentence I'd most like the youth educator to hear:

Staying out of the room doesn't protect the community. It just means somebody else picks up the pen. There's a real infrastructure problem underneath all of this, and it deserves saying plainly: AIPI's research found only about 31% of Tribal households had fixed broadband access, and the digital divide has closed slowly. Data centers now being built to power these systems carry genuine water, land, and energy questions — questions that land on rural and Tribal ground first. Those are not reasons to look away. They are precisely the issues that get decided without you if you aren't at the table. Caution keeps your hands on the wheel. That's good, and I'd never argue anyone out of it. The goal isn't to trade caution for enthusiasm — it's to trade distance for control.

6. Youth, Elders, and the Part That Isn't Optional

Your young people are already encountering this technology, and that encounter will only become more common. It is in their schoolwork, their phones, their search results, and their group chats right now. The question is not whether we can build a wall high enough to keep it out. The question is whether they encounter it with guidance, judgment, and the values of their community — or without them. A young person who learns AI inside their community learns something more than a tool. They learn that their Nation had an opinion about it. They learn which questions to ask before typing. They learn that some things don't go in the box — and why. That lesson doesn't come from the technology. It comes from the people teaching it. The same is true at the other end. Elders shouldn't be handed AI as a curiosity — they should be handed it as an accommodation: plain language, larger type, read aloud, patient, available at 2 a.m. when nobody else is. And they should be the ones consulted about what it is never permitted to touch. Both ends of the generational line. Together. That's not a nice-to-have — it's how the community's judgment actually gets transmitted along with the skill.

7. We've Been Here Before

Every generation meets a new tool that feels intimidating at first. Cars. Computers. Smartphones. The internet.

I've watched every one of those arrive since the early 1970s, and I've watched the same pattern each time: an outside promise, a period of real disruption, and then communities deciding for themselves what the thing would be allowed to be. I want to be careful here, because it would be easy — and false — to say every one of those technologies was simply good, or that adoption was always the right answer. It wasn't. Some technologies arrived with costs that landed hardest on Native communities, and some were resisted for excellent reasons. The more accurate thing, and the stronger thing, is this: Native communities have repeatedly learned, adapted to, questioned, and put new technologies to work on their own terms — deciding which to adopt, which to adapt, which to resist, and which to place under community rules. That is sovereignty. It is not a new skill. It's the skill. AI is the next tool in that line. Not a bigger threat than the ones before it, and not a smaller one. Just the next one requiring the same decision Native Nations have made over and over: what will we allow this to be here?

The Honest Close AI deserves neither blind enthusiasm nor blind fear. It deserves understanding. And for Native Nations, understanding has to begin with sovereignty. A Nation should decide what AI may be used for, what information it may touch, what is permanently off limits, who may use it, who reviews its work, and where responsibility ultimately rests. AI can carry an enormous amount of administrative and knowledge work. It can help staff write, summarize, organize, research, communicate, learn, and create. It can help a small organization accomplish what once took more people or more time than it had. But capability does not equal authority. A machine can produce an answer. It cannot grant itself the right to make a decision for a Nation. The goal isn't to put AI in charge. The goal is to make sure Native people are knowledgeable enough to keep themselves in charge. So — Learn how it works. Decide what it is allowed to do. Protect what must remain protected.

Teach your people to use it well. And keep human beings responsible for the decisions that matter. AI is not a substitute for culture, sovereignty, wisdom, or community. It is a tool. And like every important tool before it, the question was never simply what the tool can do. The question is who holds it, who sets the rules, and what they choose to build with it. Native Nations have the right to answer those questions for themselves.

To the youth educator who stopped me in my tracks: thank you. Your caution is not an obstacle to this conversation — it is the reason the conversation is worth having, and the reason it should be led by people like you rather than by outside vendors. I'm looking forward to our follow-up. I suspect I'll be doing more of the listening this time.