AI, Copyright, and the Fight for Access to Our Own Stories
There’s a growing conversation about copyright, AI, and access to knowledge. A recent blog post from Anna’s Archive caught my attention. If you’re not familiar, Anna’s Archive is a shadow library, an open-access repository that works alongside sites like Library Genesis (LibGen) and Z-Library to preserve and distribute books, academic papers, and other resources that are often locked behind paywalls. Their mission is to make knowledge freely available to all, challenging the monopoly of publishers and institutions over information.
In their latest post, they argue that copyright laws are slowing AI development, while companies in other countries — less restricted by U.S. copyright — are training AI models on massive datasets pulled from shadow libraries. Their conclusion? Copyright reform is necessary, or AI carveouts should be made to allow broader data access.
I agree that copyright laws are long overdue for reform—but for reasons that go far beyond AI. Right now, copyright locks up materials for 70 years after an author’s death. That’s an entire lifetime where critical knowledge, research, and history are legally inaccessible unless you are privileged enough to pay for access or be part of an institution that provides it.
For people like me — working to build a community-based archive that preserves Indigenous history — this system creates unnecessary barriers. So much has been written about us, but not for us. Academic papers, legal documents, and historical records are often hidden behind expensive journal subscriptions or university access policies. These paywalls keep Indigenous people from engaging with, correcting, and reclaiming the knowledge that directly concerns us.
That’s why I see libraries like Anna’s Archive as critical. They provide an alternative to institutional gatekeeping, ensuring that people — especially those outside of academic and financial privilege — can access the resources they need. Knowledge shouldn’t belong exclusively to universities, publishers, or corporations. It should be available to the communities it impacts.
For me, the real issue isn’t just whether AI can be trained more efficiently. It’s about who gets to access and shape knowledge in the first place. AI is already transforming how people learn, research, and make decisions. If we don’t fix copyright laws in a way that prioritizes community access—rather than just AI company profits — we risk reinforcing the same power structures that have kept Indigenous histories out of our hands for generations.
The AI Factor: A New Wrinkle in the Conversation
AI is reshaping how we think about knowledge, and I worry about how this plays out for Indigenous communities. Governments and corporations are already training AI on copyrighted materials behind closed doors. If copyright carve outs happen, they won’t be for us—they’ll be for “national security” or corporate interests. The same institutions that gatekeep knowledge will get special permissions, while everyday people trying to reclaim their own histories will still be locked out.
This creates an even bigger problem: what happens when AI becomes the primary way people learn about Indigenous history?
We already see it happening. Most people turn to Google or ChatGPT for quick answers. Soon, these AI tools will shape what is considered true about our histories, cultures, and identities. But if the knowledge used to train them is locked behind copyright, and only AI corporations or privileged institutions can access it, we lose control over how our history is told.
This is digital colonialism—a system where Indigenous knowledge is extracted, reformulated, and redistributed by AI models trained without our oversight. It’s not just about AI getting things wrong; it’s about who has the power to decide what history looks like.
Digital colonialism doesn’t just erase our voices. It removes our ability to challenge, correct, and contextualize our own histories. AI-generated narratives, built from incomplete or biased data, can quickly become dominant versions of history, distorting how Indigenous peoples are seen — not just by outsiders, but by our own future generations.
If we don’t intervene, we’ll wake up one day to find that the official story of our people - according to AI - wasn’t written by us.
Who Owns Our Knowledge?
Writers, historians, and researchers deserve compensation for their work, especially those writing with and for Indigenous communities. These stories matter. They help us understand who we are and where we come from. Each one is a thread in a larger tapestry, and the more we collect, the clearer the picture becomes.
But I do not believe that knowledge about my community should be locked behind paywalls or controlled by outside institutions. These stories are not just academic exercises or marketable content — they’re the voices of my ancestors. And yet, many of these materials exist because someone else had the privilege to record, archive, and publish them. That work is valuable. But access to those stories should belong to the people they come from, not just universities, publishers, or AI training datasets.
The truth is, most O’odham and other Indigenous peoples were never given that privilege. For generations, we were excluded from higher education, denied entry into universities, and systematically kept out of fields like anthropology, history, and science. Indian schools were designed to enforce manual labor over critical thinking - when our success was a goal at all.
Many of the researchers who recorded our stories cared deeply about preserving them. Some worked alongside Indigenous people who guided them, shared knowledge, and shaped the narratives being written. But academia at the time favored non-Native researchers, leaving Indigenous voices out of the final record - our voices evident in the material, but missing from the authorship.
Now, institutions that once excluded us from producing knowledge are the same ones that restrict our access to it. Indigenous knowledge sits in archives, locked behind copyright, forcing us to ask permission to learn about our own ancestors. Now, just as academia has controlled how our histories were documented, AI corporations are doing the same—harvesting data, training models, and locking away the results behind yet another layer of control.
Building Our Own AI, On Our Own Terms
We need to start thinking beyond critique — we need to build alternatives.
Indigenous communities must develop our own AI models, designed with governance, data sovereignty, and cultural protocols at their core. The Indigenous Protocol and Artificial Intelligence Workshops Position Paper lays out core principles for designing AI in a way that serves Indigenous communities rather than exploiting them.
Imagine an AI built with Indigenous governance at its core — one that:
Helps teach our past using oral histories from our elders
Supports intergenerational knowledge transfer
Connects people with their histories by making archival knowledge accessible
This isn’t just about keeping pace with technology. It’s about ensuring that AI serves our communities instead of extracting from them.
Returning to Copyright: Reform on Our Terms
This brings us back to Anna’s Archive and the broader question of copyright. There’s no doubt—copyright laws need to change. The current system benefits large institutions while restricting access to knowledge that rightfully belongs to the communities it documents.
But the solution isn’t making AI training easier for corporations. We need to open access to these archives to everyone!
Indigenous communities must be leading these discussions—not just included as an afterthought. Copyright reform should not be dictated by tech oligarchs searching for loopholes. It must be shaped by those who have been denied access to their own histories for too long.
AI, copyright, and knowledge access are all intertwined. Right now, we are at a crossroads:
If AI is allowed to be trained on copyrighted Indigenous knowledge, but we don’t have access to that same knowledge, we lose control over our history.
If AI becomes the main way people learn about Indigenous peoples, but we don’t shape those models, we lose the ability to tell our own stories.
If we don’t create AI that reflects our values, we will be forced to rely on systems that do not serve us.
The time to act is now. We need to challenge copyright laws that prevent us from accessing our own histories, push back against AI systems that exclude Indigenous voices, and build our own AI models that are governed by our own values.
The stories of our people should not be controlled by archives, universities, or AI companies.
They belong to us.
(This article written in collaboration with Chat GPT 4o. You can check out the thread HERE if interested. Final output was changed a bit, but you get the idea)