Human Rights Commission calls for human rights-centred, Te Tiriti-based approach to AI

The Human Rights Commission is calling for a human rights and Te Tiriti-centred approach to artificial intelligence and digital technologies.

The commission's new report, AI and Digital Technologies: A Human Rights and Te Tiriti o Waitangi Approach, was released on Friday and said Māori data sovereignty and Te Tiriti o Waitangi must be at the heart of how artificial intelligence was developed and governed in Aotearoa.

Rongomau Taketake Indigenous Rights governance partner Dayle Takitimu said Te Tiriti provided essential guidance for New Zealand's approach.

"Māori, like all peoples, have the right to determine how data relating to their communities, interests and identities is collected, used, governed and protected. As AI systems increasingly draw on language, knowledge and data, it is critical that Māori rights, interests and authority are respected."

Takitimu said a Te Tiriti-based approach was about "recognising tino rangatiratanga, protecting taonga and ensuring Māori are active partners in shaping Aotearoa's digital future".

"If we get this right, we can create technology that benefits everyone, while honouring our constitutional foundations."

The report said artificial intelligence could bring significant benefits, including improving access to healthcare, education and public services, supporting people with disabilities, and helping revitalise te reo Māori and indigenous knowledge.

It also identified risks including privacy breaches, discrimination and bias, digital exclusion, a lack of transparency and accountability, job displacement and the misuse of Māori data.

Senior Human Rights advisor Sophie Bradwell-Pollack told RNZ those risks were significant for Māori, because data could encompass far more than just individual information.

"It's not just ownership of basic facts about a person," she said. "It includes taonga, language, culture, and so the wider implications of that and the impact on Māori sovereignty, te ao Māori, tikanga, and all of that are a lot wider than just the kind of general kind of the smaller lens of that data is just not being used or being used in a proper way."

She said AI systems could also reproduce biases that already existed within society.

Large language models used huge amounts of data, but that data did not necessarily represent all communities equally.

"Some of the risks are that they're just not there, and not being considered and represented," she said.

"The other is also that that perpetuates biases that are already in our system, so there is a real risk there."

She said a human rights-based approach - including transparency, participation and accountability - needed to underpin how AI was used in Aotearoa.

The government is actively looking to increase the use of AI and digital tools across the public service, including as a way of improving productivity.

Bradwell-Pollack said human rights obligations did not disappear, when technology was introduced.

"Human rights and Te Tiriti obligations apply," she said. "They apply whether you're writing on paper or whether you're using a computer.

"The reality of the situation is that they also provide a tried and tested framework that we can learn from to help guide us in this kind of endeavor to figure out how we're going to govern AI for the good of everyone."

She said, if AI was used by the government to make decisions about people, there needed to be transparency around how those systems worked and what data was being used.

People also needed to know where accountability sat and how they could seek remedy, if something went wrong.

"Innovation and efficiency is only really a small part of what needs to be considered."

The commission's report said the government had primary responsibility for respecting, protecting and fulfilling human rights, and for giving effect to Te Tiriti through law, policy, investment and public-service delivery.

Bradwell-Pollack said a human rights-based approach did not prescribe a particular technology or model, but it provided a framework for ensuring people were not left without recourse, when technology affected their lives.

"These are not decisions that are being made about us, without us, essentially."

A spokesperson for Paul Goldsmith, the minister responsible for the public service and digitising government, said Goldsmith had received the report and "will consider its recommendations in due course".

"The government will table its response in parliament."

What would a Te Tiriti-based AI system look like?

Bradwell-Pollack said there was no single solution for creating a human rights and Te Tiriti-based approach to AI.

Because much of the technology was developed overseas and imported by businesses, Aotearoa could not necessarily control how the technology itself was built, but it could decide how that technology was used and governed here.

From a Te Tiriti perspective, she said that meant meaningful engagement and partnership, ensuring Māori understood how their data was being used, and supporting iwi and hapū to exercise their right to self-determination.

"It really has to be meaningful. It's not a tick-box exercise."

The report recommended a digital infrastructure strategy that looked beyond the public sector and considered how digital services could benefit the country as a whole.

Bradwell-Pollack said it would require action from government, iwi and hapū, businesses, communities and the public.

"It's not one thing. It is a multi-faceted approach that will take the government, that will take us, that will take iwi and hapū, that will take the public and business to all get on board to do."

The commission was engaging with political parties, as they developed their AI policies before the election.

The commission said Aotearoa was in a strong position to learn from approaches being developed overseas, while building on indigenous rights and Māori data sovereignty.

Bradwell-Pollack added Māori were already looked to internationally for their work in the indigenous rights and AI space.

"There is a really positive opportunity to actually put indigenous rights, data sovereignty and the obligations under Te Tiriti at the heart of what we do, and be an example to the rest of the world."

She said the technology also presented opportunities for indigenous communities, including supporting indigenous languages and helping people with disabilities, but those benefits would not happen automatically.

"These technologies exist," she said. "They are being used and, actually, there are just some things that we need to do to ensure that they're being used in a way that doesn't do harm.

"It's not that this is inevitable. We actually have a voice and we should be using that."