Reo Māori Corpora Format for generative Māori AI

Intern

Punua Waitoki, University of Waikato

Supervisor

Associate Professor Te Taka Keegan, University of Waikato

Overview

This internship investigated suitable Māori language data formats and quantities that can be best utilised by modern AI systems to build generative AI tools for te reo Māori that support Māori Data Sovereignty. The research informs investigating, cataloguing, and transforming a number of Māori language corpora that are available to us. 

Some research was undertaken to see which formats are the most suitable for our local generative AI tools. At all times the concept of kaitiakitanga over the data was maintained. The project aimed to assist Ngati Maniapoto to build a corpora of reo data that can be used in the creation of a Maniapoto reo AI.