Artificial intelligence at odds with indigenous data sovereignty

Perhaps the most sobering section of the study explores how AI can entrench harm. A key concern is data exploitation: AI relies on massive datasets, yet Indigenous communities have historically had little control over how their data is collected, stored, or used. The review emphasizes conflicts between the tech industry’s push for open data and Indigenous Data Sovereignty principles, which assert Indigenous ownership, control, and consent over data involving their peoples, lands, and knowledge.


CO-EDP, VisionRICO-EDP, VisionRI | Updated: 19-06-2025 22:22 IST | Created: 19-06-2025 22:22 IST
Artificial intelligence at odds with indigenous data sovereignty
Representative Image. Credit: ChatGPT

New research has raised the alarm over the way artificial intelligence (AI) technologies are engaging with Indigenous peoples and their knowledge systems, spotlighting both promising applications and profound systemic risks. The study, titled “Indigenous Peoples and Artificial Intelligence: A Systematic Review and Future Directions” and published in Big Data & Society, delivers the first global systematic review of literature at the intersection of Indigenous Knowledge Systems (IKS) and AI, covering publications between 2012 and 2023.

Drawing from 53 peer-reviewed sources across various disciplines and geographies, the research reveals a complex, rapidly evolving field. The authors identify four core thematic areas: AI for the preservation of IKS, AI supporting Indigenous needs, ethical and cultural concerns with AI, and the potential for IKS to reshape AI development itself. While some advances suggest AI can empower Indigenous communities through language revitalization, health care, and environmental conservation, other findings warn of entrenched algorithmic bias, data exploitation, and cultural erasure that could replicate or even worsen colonial harms.

Can AI promote and protect Indigenous knowledge?

One of the key positive trends identified is the application of AI in safeguarding Indigenous languages, lands, and traditions. The review found that AI technologies are being used to support endangered language preservation through automated transcription, translation, and text-to-speech tools. For example, speech recognition and natural language processing models have been adapted to support languages such as Cook Islands Māori and Seneca, often through transfer learning and data augmentation to compensate for limited training data. Educational apps have also emerged, such as one teaching Western Australian wildflower names in Noongar language, integrating machine learning with cultural education.

Beyond language, AI has been deployed to protect cultural and ecological heritage. In Peru and Ecuador, AI-supported gamification and interactive tools have preserved stories, customs, and historical records of Indigenous communities. Machine learning has been used to identify culturally significant plant species, such as redcedar trees vital to the Heiltsuk First Nations in Canada, and to monitor illegal deforestation in the Amazon impacting uncontacted tribes. AI-enhanced virtual reality simulations, such as those developed for Australian Aboriginal Dreaming traditions, have provided immersive educational experiences that center Indigenous worldviews.

Despite these encouraging examples, the study stresses that AI tools must align with community needs. Some Indigenous groups prioritize language documentation, while others focus on cultural resilience or environmental protection. In every case, context-specific design and Indigenous leadership in tech development are critical to ensuring AI serves community goals rather than external agendas.

How does AI impact Indigenous communities in practice?

The second category explored in the study focuses on AI applications developed to meet Indigenous challenges in health, education, and environmental management. In healthcare, AI-based diagnostic tools have been trialed to detect diabetic retinopathy in Aboriginal Australians, and predict fall risks among elderly Indigenous populations in Taiwan. AI is also being used to detect ear disease in remote Aboriginal communities and to monitor long-term wildlife data collected by Rakiura Māori in New Zealand.

In education, culturally tailored computing initiatives aim to increase Indigenous participation in science and technology. Robotics programs and storytelling-based platforms are being employed to engage Indigenous youth and enhance representation in STEM fields. Environmental applications include AI-powered water and land-use management systems, such as those used in Kenya and Guatemala to protect traditional wells and predict food insecurity. Community-led digital story-mapping for tourism initiatives also demonstrates how AI can support cultural economy development.

However, the study cautions that while AI offers tools to address these challenges, the systems must be designed with Indigenous involvement from inception to deployment. Otherwise, the risk remains that such technologies may reproduce the very inequalities they seek to resolve.

What are the risks of AI to indigenous rights and knowledge?

Perhaps the most sobering section of the study explores how AI can entrench harm. A key concern is data exploitation: AI relies on massive datasets, yet Indigenous communities have historically had little control over how their data is collected, stored, or used. The review emphasizes conflicts between the tech industry’s push for open data and Indigenous Data Sovereignty principles, which assert Indigenous ownership, control, and consent over data involving their peoples, lands, and knowledge.

Bias in AI systems is another urgent threat. In health care, models trained on non-Indigenous datasets have been shown to misdiagnose Indigenous patients at higher rates. In child protection services, predictive algorithms disproportionately flag Indigenous children for intervention based on historically biased data. In criminal justice, risk assessment tools have ignored cultural context, leading to unfair treatment in sentencing. The study stresses that such tools often reflect systemic biases embedded in data, rather than objective truths.

The research also exposes the cultural limits of mainstream AI development. Existing platforms and programming languages frequently center Eurocentric logics, which can inhibit Indigenous youth participation and misrepresent nonlinear or relational storytelling traditions. AI systems rarely reflect Indigenous epistemologies that value interconnectedness, reciprocity, and custodianship. This is not simply a technical gap - it is a structural blind spot that shapes who AI serves and how it defines knowledge.

To address these failures, the study presents Indigenous-led frameworks for ethical AI. These include guidelines that stress relational accountability, cultural specificity, governance by Indigenous protocols, and the integration of Indigenous Knowledge Systems into AI design. However, such approaches remain rare in mainstream development, and peer-reviewed literature on practical implementation is limited.

Looking forward: Generative AI and new frontiers

The study urges scholars and policymakers to examine how the rise of generative AI, capable of creating text, images, and audio, intersects with Indigenous sovereignty, ethics, and culture. Though no existing peer-reviewed studies were found on generative AI’s impact on Indigenous peoples, concerns are mounting over the unauthorized reproduction of sacred art, music, and symbols by these models.

The authors call for urgent research into how to apply Indigenous ethical frameworks to generative technologies, which may intensify epistemic violence and commodify Indigenous knowledge without consent. They also stress the need to explore the environmental costs of AI, especially in light of the disproportionate burden Indigenous communities bear from land exploitation and resource extraction for computing infrastructure.

  • FIRST PUBLISHED IN:
  • Devdiscourse
Give Feedback