AI Literacy Toolbox: AI and data sovereignty

Data Sovereignty typically refers to the understanding that data is subject to the laws of the nation within which it is stored.

Indigenous Data Sovereignty perceives data as subject to the laws of the nation from which it is collected.

Māori Data Sovereignty recognises that Māori data should be subject to Māori governance. Māori data sovereignty supports tribal sovereignty and the realisation of Maori and Iwi aspirations.

Source: Te Mana Raraunga: Māori Data Sovereignty Network. (n.d.). Home. https://www.temanararaunga.maori.nz/

Māori data governance is...about putting Māori data in Māori hands...to have control over our data...in a way that's tika, in a way that's safe, in a way that enables our flourishing.

So much data has been collected about us, over a very long period of time as part of the colonial relationship, and it's always been to service somebody else's agenda...We haven't been at the centre of it. Data governance is a reorienting of that relationship, so that we determine how that data are used. We determine how it's protected. We determine the narratives that flow from our data. Data governance gives effect to that.

Source:  Kukutai, T. (n.d.). What is Māori data governance? [Video]. Te Kāhui Raraunga. https://www.kahuiraraunga.io/#

Rangatiratanga | Authority

• Control: Māori have an inherent right to exercise control over Māori data and Māori data ecosystems. This right includes, but is not limited to, the creation, collection, access, analysis, interpretation, management, security, dissemination, use and reuse of Māori data.
• Jurisdiction: Decisions about the physical and virtual storage of Māori data shall enhance control for current and future generations. Whenever possible, Māori data shall be stored in Aotearoa New Zealand.
• Self-determination: Māori have the right to data that is relevant and empowers sustainable self-determination and effective self-governance.

Whakapapa | Relationships

• Context: All data has a whakapapa (genealogy). Accurate metadata should, at minimum, provide information about the provenance of the data, the purpose(s) for its collection, the context of its collection, and the parties involved.
• Data segregation: The ability to disaggregate Māori data increases its relevance for Māori communities and iwi. Māori data shall be collected and coded using categories that prioritise Māori needs and aspirations.
• Future use: Current decision-making over data can have long-term consequences, good and bad, for future generations of Māori. A key goal of Māori data governance should be to protect against future harm.

Whanaungatanga | Obligations

• Balancing rights: Individuals’ rights (including privacy rights), risks and benefits in relation to data need to be balanced with those of the groups of which they are a part. In some contexts, collective Māori rights will prevail over those of individuals.
• Accountabilities: Individuals and organisations responsible for the creation, collection, analysis, management, access, security or dissemination of Māori data are accountable to the communities, groups and individuals from whom the data derive.

Kotahitanga | Collective benefit

• Benefit: Data ecosystems shall be designed and function in ways that enable Māori to derive individual and collective benefit.
• Build capacity: Māori Data Sovereignty requires the development of a Māori workforce to enable the creation, collection, management, security, governance and application of data.
• Connect: Connections between Māori and other Indigenous peoples shall be supported to enable the sharing of strategies, resources and ideas in relation to data, and the attainment of common goals.

Manaakitanga | Reciprocity

• Respect: The collection, use and interpretation of data shall uphold the dignity of Māori communities, groups and individuals. Data analysis that stigmatises or blames Māori can result in collective and individual harm and should be actively avoided.
• Consent: Free, prior and informed consent (FPIC) shall underpin the collection and use of all data from or about Māori. Less defined types of consent shall be balanced by stronger governance arrangements.

Kaitiakitanga | Guardianship

• Guardianship: Māori data shall be stored and transferred in such a way that it enables and reinforces the capacity of Māori to exercise kaitiakitanga over Māori data.
• Ethics: Tikanga, kawa (protocols) and mātauranga (knowledge) shall underpin the protection, access and use of Māori data.
• Restrictions: Māori shall decide which Māori data shall be controlled (tapu) or open (noa) access.

Source: Te Mana Raraunga. (2018, October). Principles of Māori data sovereignty. https://static1.squarespace.com/static/58e9b10f9de4bb8d1fb5ebbc/t/5bda208b4ae237cd89ee16e9/1541021836126/TMR+Ma%CC%84ori+Data+Sovereignty+Principles+Oct+2018.pdf 

Issues around the control of data made available through the internet have been around before the recent developments in AI and large scale practice of data scraping, but AI has amplified them.

Background

In the context of open data, data sovereignty raises questions about who has the right to publish and share certain types of data, particularly data related to indigenous peoples, marginalized communities, or sensitive topics. There are concerns that open data initiatives may inadvertently expose or commodify data without the full consent and participation of the communities it represents. This can lead to a loss of control over how that data is used and the potential for it to be misinterpreted or used in ways that harm those communities.

Similarly, with big data, the massive scale and diverse sources of data can make it challenging to determine data ownership and ensure appropriate data governance. When large datasets are aggregated from multiple sources, it may be difficult to trace the origins of the data or ensure that it is being used ethically and in alignment with the interests of the individuals or communities it represents. This raises concerns about data sovereignty and the potential for big data analytics to perpetuate or exacerbate existing power imbalances and inequalities.

Māori data sovereignty and AI

Both open data and big data are used to train algorithms within AI models, which leads to the following issues:

• Misrepresentation and cultural appropriation: AI models trained on Māori data may perpetuate stereotypes, misrepresent Māori culture, or appropriate traditional knowledge without permission.

• Violation of indigenous rights and self-determination: The development of AI models using Māori data is a violation of the right to self-determination and sovereignty over their own data and knowledge.

• Preservation of traditional knowledge: There are concerns about the potential loss or commodification of traditional knowledge through its incorporation into AI systems.

• Lack of transparency and accountability: The opaque nature of AI systems can make it difficult for Māori to understand how their data is being used and to hold developers accountable.

• Exacerbation of existing power imbalances: The development of AI models using Māori data can reinforce existing power imbalances and colonial relationships between Māori communities and dominant societal institutions.

Created with the assistance of GenAI: Anthropic. (2024). Claude 3 haiku (March 4 version) [Large language model]. https://claude.ai/

The Waitangi Tribunal (2023) established that, as an element of mātauranga Māori, Māori data is, or has the potential to be, a taonga. Two inferences can therefore be made from this:

1. Under article two of te Tiriti, which guarantees rangatiratanga, or possession, of taonga (Orange, 2011), Māori have inarguable rights to and over any data about Māori, including data used to train AI.

2. The New Zealand government and any party required to operate with respect to the principles of Te Tiriti, have a duty to "actively protect [the] rights and interests" of Māori in the digital domain (Waitangi Tribunal, 2016, p. 50).

How can te Tiriti be upheld in the AI space?

The white paper, "Aotearoa New Zealand Artificial Intelligence: A Strategic Approach" (Bifet et al., 2021), suggests ways in which te Tiriti may be upheld and Māori data can be used in a culturally safe way in the development of AI. It asserts that ultimately there must be investment in Māori-led research to ensure sovereignty is maintained and that Māori are in control of the narrative. If this doesn't happen, we are subject to AI technology developed outside of New Zealand that doesn't value indigenous data sovereignty.

Within New Zealand, commitments to honouring Te Tiriti o Waitangi in the use of AI algorithms, like the "Algorithm Charter for Aotearoa New Zealand" (StatsNZ, 2000), set an important precedent, but as Te Kahui Raraunga (2021) argue, this cannot be fully achieved without co-governance with agencies who need use of Māori data, and that "Protocols for co-governance must empower tino rangatiratanga" (p. 12).

Research funded by Te Whatu Ora Waitemata into governance around AI used for New Zealand health services goes further, explicitly outlining how such protocols can and should be established with reference to te Tiriti and how research can be undertaken collaboratively based on tikanga and mātauranga Māori (Whittaker et al., 2023).

Protecting Māori data sovereignty outside New Zealand

Māori data sovereignty shares the same objectives and values as global efforts around indigenous data sovereignty, including the Global Indigenous Data Alliance (GIDA) (Greaves et al., 2023) and the UN Declaration on the Rights of Indigenous People (UNDRIP), which asserts the right for indigenous people to have self-determination (United Nations, 2018).

While work is being done internationally in this space, "there is no explicit international agreement governing cross-border data flows" (Aaronson, 2021). Therefore, the government needs be doing more to safeguard Māori data sovereignty, not only as a treaty partner, but as a signatory of UNDRIP. The Waitangi Tribunal (2023) recognises that "The Crown’s duty of active protection becomes ever more urgent in light of the widening reach and rapid evolution of international instruments" (p. 51) and failure to seek advice from Māori about Māori digital interests is a breach of te Tiriti.

References

Aaronson, S. A. (2021, August). Data is disruptive: How data sovereignty is challenging data governance [White Paper]. Hinrich Foundation. https://www.wita.org/wp-content/uploads/2021/08/Data-is-disruptive-Hinrich-Foundation-white-paper-Susan-Aaronson-August-2021.pdf

Bifet, A., Green, R., Wilson, D., & Zhang, M. (2021, November). Aotearoa New Zealand artificial intelligence: A strategic approach [White paper]. Artificial Intelligence Researchers Association. https://www.airesearchers.nz/site_files/28243/upload_files/AIWhitePaper.pdf?dl=1.

Greaves, L. M., Latimer, C. L., Muriwai, E., Moore, C., Li, E., Sporle, A., Clark, T. C., Milne, B. J. & (2023). Māori and the Integrated Data Infrastructure: An assessment of the data system and suggestions to realise Māori data aspirations [Te Māori me te Integrated Data Infrastructure: he aromatawai i te pūnaha raraunga me ngā marohitanga e poipoia ai ngā wawata raraunga Māori]. Journal of the Royal Society of New Zealand, 54(2), 190-206. https://doi.org/10.1080/03036758.2022.2154368

Orange, C. (2011). The Treaty of Waitangi. Bridget Williams Books.

StatsNZ. (2000, July). Algorithm charter for Aotearoa New Zealand. https://data.govt.nz/assets/data-ethics/algorithm/Algorithm-Charter-2020_Final-English-1.pdf

Te Kāhui Raraunga. (2021, March 12). Iwi data needs. https://www.kahuiraraunga.io/_files/ugd/b8e45c_4ecefb8047ab4162a3ff07468af5a27d.pdf

United Nations. (2018). United Nations declaration on the rights of indigenous peoples. https://www.un.org/development/desa/indigenouspeoples/wp-content/uploads/sites/19/2018/11/UNDRIP_E_web.pdf

Waitangi Tribunal. (2023). The report on the comprehensive and progressive agreement for trans-Pacific partnership (Report No. 2522). https://forms.justice.govt.nz/search/Documents/WT/wt_DOC_195473606/Report%20on%20the%20CPTPP%20W.pdf

Whittaker, R., Dobson, R., Jin, C. K., Style, R., Jayathissa, P., Hiini, K., Ross, K., Kawamura, K., Muir, P., the Waitematā AI Governance Group, Mark, A., Armstrong, D., Frost, E., Buxton, J., Lunny, J., Andrew, P., Bloomfield, S., Puddle, S., & Miles, W. (2023). An example of governance for AI in health services from Aotearoa New Zealand. NPJ Digital Medicine, 6(1), 1–7. https://doi.org/10.1038/s41746-023-00882-z

Pacific Data Sovereignty is understanding our ancestral heritage while upholding the ethos that ‘Nothing is about Us – Pacific Peoples – without us – Pacific Peoples.’ This ensures that the lineage of Pacific Data paves its paths and linkages back to Pacific ancestors and the generations following to present-day Pacific peoples.

It is the right of Pacific Peoples to take ownership over their own data, what happens to it, how it is collated, used, analysed, presented, and disseminated.

Pacific data sovereignty is focused on:

Source: Moana Research (2021, February). Pacific Data Sovereignty Network: Consultation document. Pacific Data Sovereignty Network. https://www.moanaresearch.co.nz/wp-content/uploads/2021/03/PDS-consultation-document.pdf