The Potential Role of Artificial Intelligence (AI) in Early Te Reo Māori Language Acquisition

Abstract

Despite 34% of Māori adults being fluent in te reo Māori, only a small proportion speak the language fluently at home with their tamariki (children). This gap highlights the challenges of language trauma, structural barriers, societal undervaluing of te reo, and broader socio-political factors related to the experience of colonisation. The importance of addressing these barriers is an urgent priority as the survival of the Māori language, central to Māori cultural identity and holistic wellbeing, is dependent on intergenerational transmission.

This thesis examines whether artificial intelligence (AI), integrating Western approaches and mātauranga Māori (Māori knowledge, education, wisdom), can enhance early te reo Māori (the Māori language) acquisition. The thesis aims to address two key questions, 1) How do Western neuroscientific understandings of the role of rhythm in language development and traditional Māori language transmission practices align? And 2) can transdisciplinary AI lend to ethical, safe, socially and culturally aligned AI for use with marginalised Indigenous, tribal, and minority (ITM) communities?

An integrative kaupapa Māori approach guides the research methodology. Through this lens, the historical, cultural, neurobiological and technical aspects of language transmission are looked at to understand how these factors might inform the development of conversational AI to augment te reo Māori uptake in the early years. These tools, when designed with cultural authenticity and in collaboration with ITM communities, offer promising pathways for creating immersive and supportive environments for language learning. However, the risks and ethical concerns surrounding AI development, particularly for Indigenous languages, are scrutinised.

Delving into the neuroscience of infant language development, the thesis seeks to show how early interaction, neural synchrony to rhythm, and traditional cultural practices such as oriori (chant composed and sung to children) and mōteatea (traditional laments or chants) align and contribute to Western understandings of infant language acquisition. These strands of knowledge could be interwoven to create a unique pedagogy for training an AI tool that optimises early uptake of native language. Drawing on research from socio and cultural linguistics, AI science, and infant neuroscience, the final chapter explores a unique Māori cultural framework, Te Wheke (the octopus), to achieve transdisciplinary AI. Transdisciplinary approaches in AI development could hold the key to addressing some of the current social, cultural and ethical concerns that hamper AI use within Indigenous language revitalisation contexts.