Indigenous Data Sovereignty and AI

What is data?

Raw, uninterpreted facts that can be either:

• Quantitative (numerical)

• Qualitative (descriptive)

What counts as Indigenous data?

Any facts, information, or knowledge created by or about Indigenous peoples, including:

• Cultural knowledge and practices

• Genealogies (whakapapa)

• Language data

• Demographic data

• Health data

Define Indigenous Data Sovereignty.

The right of Indigenous peoples to own, control, access, and make decisions about data created by or about them

Why have Indigenous peoples historically been called data warriors?

Because they have long maintained, protected, and transmitted knowledge systems across generations.

How does Te Tiriti o Waitangi support Indigenous data sovereignty?

It affirms rangatiratanga (authority and self-determination) over lands, taonga, and cultural knowledge, including non-physical forms of knowledge

What is UNDRIP's significance for Indigenous data sovereignty?

Although not legally binding, it:

• Affirms Indigenous rights

• Recognises Indigenous intellectual property rights

• Supports Indigenous control over cultural knowledge

What does Rangatiratanga mean in Māori Data Sovereignty?

Authority and decision-making power over Maori data, including where and hw it is stored and used

What does Whakapapa mean in Māori Data Sovereignty?

Recognition of relationships and origins, including understanding where data comes from and its cultural connections

How does Whakapapa relate to AI training data?

It emphasises tracing the origins of data and recognising the communities and contexts from which it was collected.

What does Whanaungatanga mean in Māori Data Sovereignty?

Obligations and responsibilities arising from relationships, balancing individual interests with collective community interests.

What responsibilities does Whanaungatanga place on data users?

People handling data have obligations to these communities from which the data was gathered

What does Kotahitanga mean in Māori Data Sovereignty?

Collective benefit—data use should create benefits for the wider Māori community.

What does Kotahitanga imply for AI development?

AI initiatives should help build community capabilities and generate collective benefits rather than only private profit.

What does Manaakitanga mean in Māori Data Sovereignty?

Reciprocity, care, and ethical responsibility in collecting and using data

What ethical requirements arise from Manaakitanga?

Data collection should involve:

• Respect

• Reciprocity

• Free, prior, and informed consent

What does Kaitiakitanga mean in Māori Data Sovereignty?

Guardianship and stewardship over Māori data.

What does Kaitiakitanga imply about data sharing?

Māori communities should determine whether data remains closed, shared, or openly accessible.

Why are LLMs often seen as incompatible with Indigenous Data Sovereignty?

Training data is often collected without clear consent, governance, attribution, or community control

How do LLMs challenge the principle of Whakapapa?

Training data is mixed and transformed, making it difficult to trace where information originated

Why might AI be incompatible with Indigenous Data Sovereignty?

AI systems often remove knowledge from its cultural context and place it into generic digital systems.

What can happen to Indigenous knowledge when used in AI?

Cultural meaning, context, and relationships may be lost or distorted.

What governance issues does AI create?

Questions arise about who controls data, who benefits, and whether Indigenous communities retain authority over their knowledge.

What is disaggregation in AI systems?

Breaking knowledge into small fragments that lose their original relationships and context

What is homogenisation in AI?

Flattening diverse knowledge into standardised forms, reducing important cultural and linguistic differences

What threat does homogenisation pose?

Unique dialects, perspectives, and cultural variations may disappear if one dominant form becomes standard.

Define data colonialism.

The appropriation and extraction of data from communities without ensuring benefits return to those communities

Why is data colonialism compared to land colonisation?

Both involve extracting valuable resources from communities while outsiders receive most of the benefits.

What is value extraction?

Turning community knowledge and data into economic value that primarily benefits external organisations.

How can AI contribute to data colonialism?

AI systems can generate profits from Indigenous knowledge without Indigenous ownership, governance, or benefit-sharing