Indigenous Knowledge and AI Governance: What Responsible AI Can Learn from the World's First Knowledge Systems

Artificial intelligence governance is often framed as a conversation between governments, technology companies, researchers, and regulators. Discussions about transparency, fairness, accountability, and safety dominate policy debates from Brussels to Washington, Beijing to Nairobi. Yet, one perspective remains strikingly underrepresented: Indigenous knowledge systems.

This absence matters. Indigenous peoples comprise more than 476 million people across over 90 countries, stewarding vast amounts of the world's cultural heritage, biodiversity, ecological knowledge, and linguistic diversity. At the same time, AI systems are increasingly trained on data, texts, images, languages, and cultural materials that originate from Indigenous communities; often without meaningful consent, participation, or benefit-sharing.

As Europe advances implementation of the EU AI Act and countries across the Global South develop their own AI strategies, indigenous perspectives offer more than a call for inclusion. They challenge some of the foundational assumptions underlying mainstream AI ethics itself.

The question is no longer whether Indigenous knowledge belongs in AI governance. The question is whether AI governance can be truly responsible without it.

Beyond Individual Rights: A Different Understanding of Data

Most contemporary AI governance frameworks are built around individual rights. Privacy laws focus on individual consent. Data protection frameworks protect individual subjects. Fairness assessments evaluate outcomes for individual users while human rights frameworks protect individual autonomy and dignity. However, while these protections are essential, indigenous perspectives reveal an important limitation: not all data is individual. Meaning knowledge about a sacred site, a traditional medicinal practice, a language, an ecosystem, or a community's history may belong collectively rather than individually.

An individual community member may not have the authority to give away that knowledge, just as one citizen cannot sell a nation's cultural heritage. This distinction lies at the heart of Indigenous Data Sovereignty i.e. the principle that Indigenous peoples have the right to govern the collection, ownership, interpretation, use, and sharing of data relating to their peoples, lands, cultures, and resources. The concept has gained traction globally through Indigenous governance movements in Canada, Australia, New Zealand, and among other countries.

For AI governance, this raises a profound challenge:

Can consent from individuals alone be sufficient when the knowledge being extracted is collective?

How do we transition from FAIR to CARE?

One of the most influential interventions from Indigenous scholars has been the development of the CARE Principles for Indigenous Data Governance. For years, data governance was guided primarily by the FAIR principles: data should be Findable, Accessible, Interoperable, and Reusable. While valuable, FAIR largely focuses on technology accessibility, indigenous scholars argued that accessibility alone does not address questions of power, justice, or self-determination.

In response, the Global Indigenous Data Alliance developed CARE, emphasizing:

• Collective Benefit
• Authority to Control
• Responsibility
• Ethics

This shift is significant. CARE asks not only whether data can be used, but whether communities benefit from its use, who has authority to govern it, and what responsibilities accompany access.

For AI developers, policymakers, and researchers, this changes the conversation from technical compliance to governance. As such, responsible innovation is not only about reducing harm. It is also about ensuring that communities retain meaningful agency over how technologies are developed and deployed.

When Data Extraction Becomes Digital Colonialism

Many communities in the Global South will find these concerns familiar. For centuries, colonial systems extracted resources, labor, knowledge, and wealth from colonized societies while concentrating power elsewhere. Today, similar dynamics can emerge in digital environments. For example, data generated in African, Asian, Latin American, Pacific, and indigenous communities often flows into AI systems developed elsewhere i.e., located mainly in the Global North. The resulting technologies largely generate enormous economic value, yet local communities rarely participate in governance decisions or share proportionately in the benefits.

Researchers increasingly describe this phenomenon as data colonialism or digital colonialism. Indigenous scholars have warned that AI development can reproduce extractive relationships when cultural knowledge, languages, and community-generated information are collected and commercialized without meaningful participation from the communities themselves. For Global South policymakers, indigenous perspectives offer valuable frameworks for resisting these dynamics; not through technological isolation, but through governance models grounded in sovereignty, reciprocity, and community benefit.

Indigenous Governance Is Already Happening

Discussions about Indigenous perspectives in AI are often framed as future possibilities. In reality, indigenous communities are already developing governance frameworks that challenge conventional approaches to data, technology, and innovation. Three cases stand out for the depth and distinctiveness of their contributions.

New Zealand: Māori Data Sovereignty and Algorithmic Accountability

Perhaps one of the most developed examples comes from New Zealand, where Māori organizations have spent years advancing indigenous data sovereignty. The Māori Data Sovereignty Network, Te Mana Raraunga, argues that Māori data should be subject to Māori governance and that Indigenous communities should have authority over how their data is collected, used, and shared. Their work challenges a common assumption in AI development: that data ownership is primarily an individual matter. Instead, Māori scholars emphasize collective rights, cultural stewardship, and community benefit.

This perspective has become increasingly relevant as governments deploy automated systems in public services. Questions around predictive analytics, welfare systems, and algorithmic decision-making are not only about accuracy or fairness; they are also about who has authority over data and whose values are embedded in technological systems.

New Zealand's Algorithm Charter for Aotearoa New Zealand provides one example of how governments can incorporate accountability principles into public-sector algorithmic governance.

Canada: Indigenous Data Governance and the OCAP® Principles

In Canada, First Nations communities developed one of the world's most influential Indigenous data governance frameworks: the OCAP® Principles (Ownership, Control, Access and Possession). OCAP establishes that First Nations communities; not external organizations; should determine how information about their communities is managed and used.

While OCAP predates today's generative AI boom, its principles have become increasingly relevant as large language models are trained on vast quantities of text, cultural materials, and community-generated information.

For AI governance, OCAP raises important questions:

• Can publicly available Indigenous knowledge be used for AI training without community consent?
• Should communities benefit when their cultural resources contribute to commercial AI products?
• How should Indigenous languages be protected while also being supported by AI technologies?

Canada's experience demonstrates that governance is not confined to protecting privacy; it equally involves protecting community authority.​

Australia: Indigenous Protocols for AI and Data

Australia has become an important testing ground for discussions about Indigenous knowledge, AI, and digital governance. Organizations such as the Maiam nayri Wingara Indigenous Data Sovereignty Collective have argued that Indigenous peoples must have meaningful control over data concerning their communities, territories, and cultures. This work is particularly important as AI systems increasingly rely on geospatial information, environmental data, and cultural archives that may contain Indigenous knowledge.

Australia's experience demonstrates that responsible AI governance is not only about preventing harm. It is also about recognizing Indigenous peoples as rights-holders and decision-makers rather than stakeholders consulted after decisions have already been made. The AIATSIS Code of Ethics for Aboriginal and Torres Strait Islander Research offers valuable lessons for community-led governance and informed consent.

Finland, Norway, and Sweden: Sámi Rights in the Digital Age

The Indigenous Sámi people span Norway, Sweden, Finland, and Russia, offering an important European perspective on Indigenous digital rights. Across the Nordic region, debates have emerged around language preservation, cultural heritage digitization, and data governance. As AI-powered language technologies become more common, there are growing concerns about how Indigenous languages are collected, digitized, and used.

Institutions such as the Sámi Parliament of Finland and the Sámi Council have helped advance conversations about indigenous self-determination and cultural rights. For Europe, the Sámi experience highlights an often-overlooked dimension of AI governance: cultural rights. Responsible AI cannot solely protect privacy. It must also consider linguistic diversity, cultural continuity, and community self-determination.

Brazil and the Amazon: AI, Environmental Governance, and Indigenous Stewardship

In Brazil, indigenous communities are increasingly involved in digital monitoring and environmental governance initiatives across the Amazon. AI and satellite technologies are being used to track illegal mining, deforestation, and land encroachment. Yet these initiatives also raise questions about who controls environmental data and who benefits from technological interventions.

Organizations such as the Coordination of Indigenous Organizations of the Brazilian Amazon (COIAB) have consistently argued that environmental governance systems should not merely extract information from Indigenous territories but should empower Indigenous communities to lead decision-making processes. This lesson extends beyond Brazil. Across the Global South, AI systems designed for climate resilience, agriculture, conservation, and resource management must recognize Indigenous peoples as knowledge holders rather than data sources.

What Africa Can Learn

Africa is home to thousands of Indigenous and traditional knowledge systems, many of which remain underrepresented in global AI discussions. South Africa provides one of the continent's most developed examples of formal recognition for Indigenous knowledge. Through its Indigenous Knowledge Systems Office and broader Indigenous Knowledge Systems policy framework, the country has sought to protect, promote, and commercialize Indigenous knowledge in ways that benefit local communities rather than external actors.

This recognition emerged partly from concerns that valuable Indigenous knowledge; including traditional medicine, agricultural practices, biodiversity knowledge, and cultural heritage; was being documented, researched, and commercialized without adequate acknowledgment or benefit-sharing for the communities that developed and preserved it.

These concerns are increasingly relevant in the age of AI. As AI systems are trained on larger and more diverse datasets, traditional knowledge can become a source of valuable information for applications ranging from healthcare and climate adaptation to agriculture and biodiversity management. Yet without appropriate governance mechanisms, AI developers could extract and utilize this knowledge while communities receive little influence over how it is used or who benefits from it. South Africa's approach offers an important lesson: knowledge should not be viewed solely as a resource to be collected and monetized. It should also be understood as a form of cultural heritage and community asset that requires stewardship, protection, and equitable governance.

The implications extend beyond South Africa. Across the continent, governments are developing national AI strategies and digital transformation agendas. Countries such as Kenya, Nigeria, Rwanda, and Ghana are exploring how AI can support economic development, public services, agriculture, and innovation.

As these strategies evolve, Indigenous governance frameworks raise critical questions:

• Who owns traditional and community-held knowledge that may be used to train AI systems?
• How can communities provide meaningful consent when knowledge is collectively rather than individually held?
• What mechanisms ensure that communities share in the benefits generated by AI systems built using their knowledge?
• How can governments protect local languages and cultural heritage while encouraging innovation?

These questions are particularly important in Africa, where many communities maintain strong traditions of collective stewardship, customary governance, and communal ownership that do not always align neatly with Western legal concepts of individual ownership and consent. Indigenous governance frameworks therefore offer more than a cultural perspective. They provide practical tools for designing AI governance systems that are rooted in local realities, protect community interests, and advance digital sovereignty across the continent.

What This Means for the EU AI Act and the Global South

The European Union's AI Act is often described as the world's most influential AI regulation. Through its risk-based approach, transparency requirements, and protections for fundamental rights, it is already shaping conversations about AI governance far beyond Europe. For policymakers in Africa, Latin America, and Asia, the AI Act offers valuable lessons. It demonstrates that governments can place guardrails around AI systems rather than leaving governance entirely to technology companies. Notwithstanding this benefit, indigenous governance frameworks reveal an important limitation: good regulation is not necessarily universal regulation.

The AI Act is rooted in European legal traditions that emphasize individual rights, consumer protection, and regulatory oversight. These are important foundations. However, many indigenous governance systems begin from different assumptions; centering collective rights, stewardship responsibilities, community authority, and obligations to future generations. As a result, a system could comply fully with the EU AI Act and still fail to address indigenous concerns. In addition, an AI developer may satisfy transparency requirements while using cultural knowledge that a community believes should not be commercialized. Furthermore, a model may comply with data protection rules while training on indigenous languages without community participation in governance decisions. Moreover, a risk assessment may evaluate harms to individual users while overlooking impacts on collective cultural heritage.

This is particularly relevant for countries in the Global South that are currently exploring AI regulation. Many governments are looking to the EU AI Act as a template, however, the goal should not be regulatory replication. The goal should be governance that reflects local realities. Just as legal systems cannot be copied wholesale from one society to another, AI governance frameworks must reflect local histories, institutions, cultures, and social priorities. The most important lesson from Indigenous governance may therefore be that responsible AI cannot be achieved through compliance alone. It requires legitimacy; and legitimacy comes from the meaningful participation and authority of the communities most affected by technological change.

A New Vision for Responsible AI

For much of the AI era, governance discussions have been dominated by governments, technology companies, and technical experts. Indigenous communities remind us that another perspective is possible. Their contribution extends far beyond a request for representation. It is a challenge to some of the core assumptions that underpin contemporary AI governance: who owns knowledge, who benefits from innovation, who has the authority to make decisions, and what responsibilities we owe to future generations.

For Europe, these perspectives highlight the limits of relying solely on compliance-based approaches to AI governance. For the Global South, they offer practical lessons on digital sovereignty, community participation, and resisting extractive models of technological development. And for the global AI community, they provide a powerful reminder that responsible AI is not only about managing risk. It is about governing knowledge, power, and relationships in ways that are just and equitable.

The future of AI governance should not be built exclusively in Brussels, Washington, Beijing, or Silicon Valley. It should also be informed by the governance traditions of Māori communities in Aotearoa New Zealand, First Nations peoples in Canada, Sámi institutions across Northern Europe, Indigenous peoples of the Amazon, and countless communities across Africa, Asia, and the Pacific.

If AI is to serve humanity, then humanity's diverse ways of understanding knowledge, responsibility, and collective wellbeing must have a place in shaping its future. Because the challenge facing AI governance today is not merely how to regulate powerful technologies. It is how to ensure that the communities whose knowledge, languages, cultures, and histories are increasingly shaping AI systems have a meaningful voice in determining how those systems are governed.

​That is not merely a question of inclusion. It is a question of justice.