Te Kete o Karaitiana Taiuru (Blog)

6 Te Tiriti Based Artificial Intelligence Ethical Principles by Karaitiana Taiuru

These peer reviewed principles were originally written by Dr Karaitiana Taiuru, to provide a Māori and Te Tiriti perspective on Artificial Intelligence with Health, complimenting the ‘Principals section’ written by Professor James Maclaurin – Te Whare Wānanga o Ōtākou | The University of Otago, for inclusion into the collaborative and peer reviewed report “Capturing the benefits of AI in the healthcare for Aotearoa New Zealand. A rapid response from the Prime Minister’s Chief Science Advisor, Kaitohutohu Pūtaia Matua ki te Pirimia”.

The principles have been modified to be a general set of principles for all AI. Various specialised AI such as languages, legal, education, etc will need to modify these principles to make them applicable to the various AI.

 

INTRODUCTION

The rise in the widespread usage and access of Artificial Intelligence, especially with Large Language Learning Models (LLMs) such as OpenAI and by companies such as Google and Microsoft in the past year have by far superseded any Māori Data guides and frameworks. These guidelines also reinforce the idea of creating a Te Tiriti based Māori LLM for educational purposes and to balance the current bias, censure and false media against Māori, as Latima, nicknamed the Black GPT.

The rapid growth has also raised the questions about AI regulation in New Zealand and the idea of New Zealand sovereignty for AI to protect the interests of New Zealand.

These principles are the foundational ethical principles for any AI system that uses Māori Data, produces Māori Data, or makes decisions about Māori collectively or individually. They should be read in conjunction with Māori Data Sovereignty principles.

Māori Data, as defined by the Waitangi Tribunal’s WAI 2522 Report on the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP) 2023 report as “Digital or digitisable information or knowledge that is about or from Māori people, language, culture, resources, or environments”[iv].

Māori Data Sovereignty as defined by the Waitangi Tribunal’s WAI 2522 Report on the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP) 2023 report as “Māori Data Governance. The principles, structures, accountability mechanisms, legal instruments, and policies through which Māori exercise control over Māori data”[v].

 

 

LEGAL INSTRUMENTS TO CONSIDER WITH MĀORI DATA AND ARTIFICAL INTELLIGENCE

Since 2016 there have been significant advancements for the protection and recognition of Māori Data Sovereignty, legal developments, court judgements, and Waitangi Tribunal findings that have provided Māori with greater consideration and protection with technologies.

 

New Zealand Legislation recognising Te Tiriti with Digital Tech

Legislation in New Zealand has begun to add recognition to Te Tiriti, recognising that digital and data are Taonga including:

  • Digital Identity Services Trust Framework Act 2023 – Section 8A Tiriti o Waitangi/Treaty of Waitangi
  • Data and Statistics Act 2022 – Section 14 Duties of Statistician relating to Tiriti o Waitangi/Treaty of Waitangi; Section 15 Principles of engagement by Statistician with Māori
  • Privacy Act 2020 Section 21 (c) Commissioner to have regard to certain matters.

 

Te Tiriti and He Whakaputanga

Te Tiriti o Waitangi is a founding document of government in Aotearoa New Zealand. Although it is important to note that matters of sovereignty had been addressed by the earlier signing of He Whakaputanga Rangatiratanga o Nu Tireni/ The Declaration of the Independence of New Zealand 1835. The Waitangi Tribunal asserts that this sovereignty was not erased or superseded by Te Tiriti o Waitangi.

Te Tiriti and its principles require consideration on an ongoing basis as the breadth of applications for AI creation and delivery continues to evolve.

Ko Aotearoa Tēnei: Report on the Wai 262 (2011) commonly known as the Māori Intellectual Property Rights claim of 2011, is the genesis of modern-day Māori Data Sovereignty. Despite this, it has not been acknowledged by most, despite it being responsible for:

  • Māori advisory committee for Patents (Patents MAC) are appointed by the Commissioner of Patents under section 225 of the Patents Act 2013.
  • And with the Wai 2522 claim and the WAI 262 Claim the Māori committee for plant varieties under the Plant Variety Rights Act 2022.
  • In 2019, the New Zealand government committed to’ Te Pae Tawhiti’ The whole-of-government response to the WAI 262.

 

 

Te Pou Matakana Limited v Attorney-General (No 1) [2021] NZHC 3319 (WOCA 2)

The High Court decision regarding Māori Data and Privacy in Te Pou Matakana Limited v Attorney-General [2021] NZHC 3319. Recognised collective rights to Māori Data that include non-Iwi entities and non Te Tiriti partners.

The Privacy Commissioner has a summary of both hearings.

 

New Zealand Supreme Court

These principles recognise the judgement by New Zealand’s highest ruling court – The Supreme Court, in its judgement Peter Hugh McGregor Ellis v R [2022] NZSC 115, 07 October 2022. In its judgement, it is stated that tikanga Māori (Māori customary lore) is and was New Zealand’s first common law, and that tikanga Māori can be applied individually and collectively to both Māori and non-Māori in New Zealand.

 

United Nations Declaration of the Rights of Indigenous Peoples (UNDRIP)

In addition, these ethical principles also recognise that globally, collective rights for Indigenous Peoples are recognised and affirmed by the United Nations Declaration of the Rights of Indigenous Peoples (UNDRIP). New Zealand adopted the declaration in 2010, acknowledging Māori as tangata whenua and affirming a commitment to the common objectives of the declaration and Te Tiriti o Waitangi and He Whakaputanga.

 

Legal Personhood

New Zealand has granted legal personhood to three geographic features. When considering traditional tikanga Māori (customary lore) that is recognised by the Supreme Court as New Zealand’s first common law, it is feasible that an AI or Māori Data could claim personhood.

(a) 2014. Te Urewera Act

(b) 2017. Te Awa Tupua (Whanganui River Claims Settlement) Bill

(c) 2017. Record of Understanding Mount Taranaki

These principles consider and are compatible with legal personhood considerations of AI and Māori Data.

 

PRINCIPALS

These principals have been written to consider all legal instruments that have influenced Māori Data Sovereignty rights and international data protection policies.

 

Principle 1: Tino Rangatiratanga

All AI systems will embed Māori leadership, decision-making and governance at all levels of the systems life cycle from inception, design, release to monitoring.

 

Principle 2: Equity

AI systems will achieve equity outcomes for Māori (individuals and collectively) across their life course and contribute to Māori development. This involves businesses and employees who are accountable to Māori in how AI models are used with Māori data and outputs that impact Māori individually and collectively and in the active building of capacity of the Māori AI workforce.

 

Principle 3: Active protection

Requires AI developers and product owners to act to the fullest extent practicable, to achieve equitable outcomes for Māori. This includes ensuring that it, its agents are well-informed on the extent, and nature, of both Māori outcomes and efforts to achieve Māori equity.

Ensuring Free, prior, and informed (FPIC) consent is required for the use of Māori data in AI development, with robust procedures in place to prevent biases or predictions that stigmatise or harm Māori.

 

Principle 4: Mana whakahaere

Effective and appropriate stewardship or kaitiakitanga over AI systems is required. It is recognised that Māori Data is a Taonga and subject to Māori Data Sovereignty principles determined by Te Tiriti. A deep understanding of the source and intended use of data is required, so that it is not repurposed without permission or in a way that will diminish the mana of Māori.

Shared and recognised Intelectual Property rights are proportioned to Māori who contribute Māori Data to the AI.

 

Principle 5: Mana Motuhake

Requires that tikanga are followed throughout the whole AI development and deployment cycles, with Māori deciding what data and data uses are controlled or allowed.

 

Principle 6:Tapu/Noa; Cultural safe practices.

No AI will be culturally unsafe or break the rules of Tapu and Noa.

Tapu is the strongest force in Māori life. It has numerous meanings and references. Tapu can be interpreted as ‘sacred’, or defined as ‘spiritual restriction’, containing a strong imposition of rules and prohibitions. A person, object or place that is tapu may not be touched or, in some cases, not even approached.

Noa is the opposite of tapu and includes the concept of ‘common’. It lifts the ‘tapu’ from the person or the object. Noa also has the concept of a blessing in that it can lift the rules and restrictions of tapu.

To associate something that is extremely tapu with something that is noa is offensive to Māori.

Mead’s Tikanga Test is a framework using Tikanga Māori and Mātauranga Māori to assess contentious issues to find a traditional Māori position on these issues and to assist avoiding possible breaches of tapu and to ascertain the risks.

 

Test 1:

The Tapu Aspect – Tapu relates to the sacredness of the person.  When evaluating ethical issues, it is important to consider whether there will be a breach of tapu, if there is, will the gain or outcome from the breach be worth it.

Test 2:

The Mauri Aspect – Mauri refers to the life essence of a person or object. In an ethical context, one must consider whether the Mauri of an object or a thing will be compromised and to what extent.

Test 3:

The Take-utu-ea aspect – Take (Issue) Utu (Cost) Ea (Resolution). Take-utu-ea refers to an issue that requires resolution. Once an issue or conflict has been identified, the utu refers to a mutually agreed upon cost or action that must be undertaken to restore the issue and resolve it.

Test 4:

The Precedent aspect.  This refers to looking back at previous examples of similar issues that have been resolved in the past. Precedent is used to determine appropriate action for now.

Test 5:

The principles aspect. This refers to a collection of other Māori principles or values that may enhance and inform an ethical debate including the Community-Up Model; Māori Data Ethical Framework; Compendium of Māori Data Sovereignty.

DISCLAIMER: This post is the personal opinion of Dr Karaitiana Taiuru and is not reflective of the opinions of any organisation that Dr Karaitiana Taiuru is a member of or associates with, unless explicitly stated otherwise.

3 responses to “6 Te Tiriti Based Artificial Intelligence Ethical Principles by Karaitiana Taiuru”

  1. Marie Hibbs Avatar
    Marie Hibbs

    Kia ora Karaitiana,
    Ko Marie Hibbs taku ingoa.
    I am an RTLB at Te Huinga Raukura ki Manurewa RTLB Cluster 12.
    I currently work in Te Putahi Māori o Manurewa (Tau 1-13), Te Huringa (Rumaki Reo – Year 0 – 8 Finlayson Park School) as well as with Whānau Reo Rua.
    Yesterday I heard your presentation on the AIEDCOP webinar and during that time I began to realise at a very beginning level, the importance of building our knowledge and understanding of how AI will (possibly already is) impacting Māori in Education. As you can see, I have accessed this article on the 6 Tiriti based AI ethical principles. I will be sharing my findings with our Tuia Te Muka Tangata Focus Group on Thursday and hopefully we can establish a working party to discuss our findings, especially in relation to our practice as RTLB.
    This article will open our minds and help us to find ways to incorporate our learning into our practice.
    Engari, he mihi maioha tēnei ki a koe.
    Marie Hibbs

Leave a Reply

Archive