Technologies a boost for te reo Māori learning

For many people, learning te reo Māori comes with a fear of getting it wrong. Pronunciation is often cited as one of the biggest challenges, and mistakes can lead to embarrassment or hesitation to use the language at all. When people are confident in saying words correctly, it encourages them to use te reo more often in their daily lives.

At the University of Auckland, researchers are working on pioneering projects that use artificial intelligence to support the revitalisation of te reo Māori.

Initiatives ranging from a pronunciation coach that gives real-time feedback to a system that can recognise emotions in speech are reshaping the way the language can be learned and experienced.

Dr Piata Allen (Ngāti Kahungunu, Ngāi Tūhoe), from the Faculty of Arts and Education, has seen first-hand when people feel ‘pronunciation embarrassment’ through her work teaching te reo Māori. She ran the online pronunciation course at the University for staff and teacher trainees which provided practical tools for saying words correctly. That experience laid the groundwork for something bigger.

“Pronunciation remains one of the biggest barriers for beginners,” she says.

Now Allen is co-leading a three-year project to develop an AI pronunciation coach for te reo Māori, funded by a $1 million Smart Ideas grant. Working alongside Dr Jesin James and Professor Catherine Watson (both from the Faculty of Engineering and Design), she is building a tool that will provide real-time, personalised feedback to learners.

“This tool will support accurate pronunciation in a way that’s empowering and supportive,” she says.

The project brings together Māori language experts, linguists, engineers and computer scientists. The aim is to help people develop muscle memory for pronunciation and to enable everyone working with Māori names and communities to speak confidently.

For Dr James, the motivation came from observing how existing technologies often overlook smaller languages.

“Most technology we see around us works for American English, British English, or Mandarin,” she says. “If your language is not one of those, it’s likely you’re left out of the digital world.”

The pronunciation coach is designed to go further than simply telling learners whether a word is right or wrong. It can highlight how a word was mispronounced and guide users on mouth shape and tongue position, making the learning process more precise and supportive. What makes the project unique, the team notes, is the richness of perspectives that have shaped it. With the involvement of Te Hiku Media, an iwi-owned technology company, tikanga Māori is woven through the design.

Professor Watson, an expert in speech technology, describes the coach as a way of unlocking progress for learners. Instant feedback can make all the difference when people are trying to get the sound right.

A long-term possibility is to support other endangered and minority languages around the world.

Read the full story.

While pronunciation is one part of language learning, another project is focusing on the emotions carried in te reo Māori. PhD candidate Himashi Rathnayake is leading the development of the world’s first speech emotion recognition system designed specifically for the language. The goal is to understand spoken words and also to accurately interpret the emotions behind them.

Rathnayake reflects on her own background.

“My native language, Sinhala, is also quite underrepresented in technology development,” she says. “During my studies in Sri Lanka, I learned about artificial intelligence and noticed that most AI technologies only focus on a few major languages. Others, like Sinhala and New Zealand’s Indigenous language, te reo Māori, were often left out of the digital world. That inspired me to find solutions that help make AI more inclusive for everyone.”

To build the system, the team began with focus groups that identified more than 200 Māori words for emotions. These were divided into 16 cultural categories. Words such as pōuri may convey sadness, mourning, darkness or remorse; harikoa might express happiness, delight or interest; and hōhā can indicate irritation, boredom or fatigue. Professional Māori voice actors then recorded expressive sentences to form the database used to train the system.

For Rathnayake, the stakes are high.

“If a system thinks you’re angry when you’re not, and if it’s because of your language or culture or where you come from, it’s not just inaccurate, it’s unfair,” she says.

The project is guided by the principle that emotion recognition must reflect Māori cultural understandings of how feelings are expressed.

Innovation rooted in community

Both projects grew out of practical needs. In te reo Māori courses, lecturers were spending significant time giving detailed feedback on pronunciation, a demand that inspired the idea of an automated coach. For the emotion recognition work, Māori community members reviewed media content and completed questionnaires to identify emotions in context, ensuring the categories were accurate and meaningful.

James sums up the collaborative approach: “Working with people from different disciplines and worldviews has been the most rewarding part.”

These two strands of work, pronunciation and emotion recognition, are both about voice. One helps people say names correctly, master tricky sounds and gain confidence. The other ensures that the emotions in Māori speech are not overlooked, and that technology can recognise what is being expressed not only in words but in feeling.

Rathnayake believes that "technology development for a community should be grounded in the community, carried out in partnership with the community, and ultimately serve the community’s interests".

“This work isn’t just for Aotearoa. It’s a call to technology developers around the world to build technology that’s not only smart, but also culturally sensitive,” says Rathnayake. “When machines learn to hear us, they should hear all of us.”

Read the full story