Ending with impact: Artificial intelligence to empower iwi to manage their data their way

SfTI researchers are using machine learning technologies to create a digital platform that allows whānau, hapū and iwi to interact with their data in their language and on their own terms. By developing artificial intelligence which interacts in te reo Māori, they will empower iwi across Aotearoa to build their own digital content infrastructures which are grounded in te ao Māori.

This project builds on the work of the $4.2M Spearhead project Ātea, which successfully created an immersive experience of the wharerau Tahu Pōtiki (Te Rau Aroha Marae in Motupōhue/Bluff) that enabled whānau to connect remotely back to their own stories and mātauranga. This project will expand on the Ātea web portal and develop ‘Your Ātea Natural Language Processing (NLP) Platform’, a te reo speaking virtual machine-based software solution capable of being run through a third-party infrastructure, like Google cloud.

Image: The Ātea web portal

The project is part of a new suite of SfTI funding to enable our most promising Spearhead research projects to ‘end with impact’. All National Science Challenges, including SfTI, will draw to a close in June 2024 and it’s vital that every drop of practical impact from our research is delivered. In a 2022 funding round, our Spearhead projects were invited to put forward ideas that have germinated during their work of the past few years, and that with more time and investment could reap impactful rewards. This project was one of two ‘ending with impact’ projects that have come from the Ātea Spearhead. The other project is developing a virtual/ immersive wānanga

Developing artificial intelligence in te reo

This project is called ‘Your Ātea Natural language processing Platform’. Natural language processing (NLP) is an artificial intelligence that gives computers the ability to understand text and spoken words like a human. NLP combines computational linguistics, rule-based modelling of human language, with statistical, machine learning, and deep learning models.

David Bainbridge co-leader of this project based at University of Waikato says that like most technology NLP has been primarily developed with dominant languages in mind and that does not include te reo Māori. 

“Although a huge amount of content management environments exist, they’re built only with the big population languages and primarily English in mind. Endangered languages like te reo Māori are not supported in these tools. This research project is focussed on adjusting this imbalance.”  

Massey based co-lead, Hēmi Whaanga (Ngāti Kahungunu, Ngāi Tahu, Ngāti Mamoe, Waitaha), says that many iwi want to take control and exert autonomy over their digital information, but lack the resources and expertise to do so.

"Current NLP Tools for te reo Māori content are disjointed, but some inroads have been made by the original Ātea Spearhead including tools to macronise text, generate text from speech and speech from text, spell check text, and auto text prediction. 

“We want to bring these pieces together so there is a freely available set of Natural Language Processing tools for te reo. This is a by Māori for Māori approach. ‘Nā tō rourou, nā taku rourou, ka ora ai te iwi  - With your food basket and my food basket the people will thrive’". Says Hēmi.

Mahi Ngātahi (Co-design) 

The first stage of this project is Mahi Ngātahi (Co-design). This means the project will make the most of the relationships formed as part of the original spearhead in order to enable Māori collaborators to identify how they want this technological innovation to work for their benefit. 

The Mahi Ngātahi approach will help partnering Māori communities define what impact looks like and put it into action. Kaupapa Māori community practitioners, Tū Māia, will facilitate the Mahi Ngātahi phase by undertaking a wānanga with each stakeholder and using a tool called a Kaupapa Canvas to help determine their dreams and visions for the tech. The research team will also be present at these wānanga to ensure integration into the technology itself. 

The project is funded for $1,264,907 to March 2024. Find out more about our ‘ending with impact’ projects.