New Zealand AI researcher Punahamoa ‘Puna’ Walker is building software that can speak to Indigenous communities in their language
Anyone who’s tried to learn a language knows how tricky cultural nuances can be. So what if you’re trying to teach a robot? Chatbots and the like are becoming part of daily life, but if they only speak English, lots of users will be left behind.
Artificial intelligence (AI) researcher Puna is hoping to change that. He works for FranklyAI,
a New Zealand chatbot-like platform that can speak Te Reo Māori and its iwi (or tribal) dialects. It also converses in Samoan, Mandarin and Torres Strait Creole, with more languages to come, and is used by governments, universities and other organisations
to consult with students, customers and the public.
“One of the biggest barriers to getting people involved is language and culture,” says Puna. “We want to make the experience as good as possible, like you’re talking to a real person. Different iwi dialects use different spellings or prefer different words. It means we need to put more effort into how we train it.”
Transferable skills
In some cases, there are words that don’t translate well into other languages and generational differences.
“We did a project for the University of Auckland, but we realised our translators don’t necessarily know how an 18-year old-speaks,” he says. “We had to find students who were fluent in the languages to help us.”
Puna originally studied electrical engineering and was pursuing a PhD in the US when the pandemic hit. He returned to New Zealand with a master’s instead, but it was hard to find a job that matched his research – specifically, looking at how certain sounds can aid in health problems like seizures and dementia. Fortunately, he had transferable skills.
“Being in electrical engineering for so long, you do pick up some software and coding skills,” he says.
The best part of Puna’s job? Learning about other languages and cultures, as well as his own.
“I grew up in Auckland, away from my iwi (Whakatōhea), so this has been a good opportunity to reconnect with my people and my language. It’s been a bit of a journey of self-discovery!”
Puna’s study and career path to becoming an AI researcher
- Bachelor of Engineering (Honours) (Electrical and Electronic Engineering), University of Auckland
- Product Development Intern, Fisher & Paykel Healthcare
- Master of Engineering (Electrical and Electronic Engineering), University of Auckland
- Signal Processing Engineer, Hyland Consultants Ltd
- Master of Science (Electrical and Electronic Engineering), Georgia Institute of Technology
- AI Researcher, FranklyAI
This profile was originally published in Careers with STEM: Technology
