I didn't plan a career in technology. I studied computer science at NUS in Samoa because I was good at maths and logic, and because my family valued education. When I graduated, I went into health data, then community development, then programme coordination. Technology was always the tool, never the identity. But looking back across 15 years of work, from Apia to Auckland, I've been in tech the whole time. I just didn't always see other Pacific women there with me.
Why This Matters Now
- Pacific peoples make up 8.1% of New Zealand's population but are significantly underrepresented in the technology workforce, particularly in senior and technical roles.
- AI systems are making decisions about Pacific communities, health screening, education placement, social service allocation, immigration processing. The people building these systems rarely include Pacific voices.
- Representation isn't about optics. When AI systems are built without Pacific perspectives, they reproduce the biases already present in the data, and nobody in the room knows enough to catch it.
- Pacific women bring specific knowledge about community, governance, collective decision-making, and holistic wellbeing that AI development desperately needs.
My Path Through Tech
My career doesn't look like a typical tech trajectory, which is part of the point.
After my BSc at the National University of Samoa, I worked with the National Health Service building data systems for disease surveillance and maternal health monitoring. That was technology work - designing databases, writing queries, building reporting frameworks - but nobody called it "tech." It was health sector work that happened to involve computers.
In 2012, Cyclone Evan hit Samoa. I volunteered with the Red Cross. The coordination work was data-intensive - tracking displaced populations, managing supply distribution, reconciling records across damaged infrastructure. Technology under pressure. Still not called "tech."
I moved to Auckland. At Foundation North, I coordinated grants - which meant working with data systems, applicant tracking, reporting frameworks. At Designertech, I coordinated projects - which meant working with project management tools, client databases, and delivery pipelines. At Health NZ, I coordinated health programmes - which meant working with health information systems, population data, and outcome tracking.
Each of these roles used technology daily. None of them would show up in a "women in tech" survey.
This is the first gap. Pacific women are doing technology work across health, community development, education, and government. But because they're not in roles with "engineer" or "developer" in the title, they're invisible in tech workforce statistics.
Who Was in the Room
At almost every technology meeting I've attended in Aotearoa, I could count the Pacific people on one hand. Often I didn't need any fingers.
This isn't a new observation. Pacific communities have been underrepresented in New Zealand's technology sector for as long as there's been a technology sector. But with AI, the stakes are different.
When a website is built without Pacific input, the worst outcome is that it doesn't work well for Pacific users. Annoying, but not life-altering.
When an AI system is built without Pacific input, the outcomes can be much more serious. Health screening algorithms that don't account for Pacific body composition. Risk assessment tools trained on data that reflects existing biases against Pacific communities. Automated decision systems for government services that don't understand Pacific family structures.
These aren't hypothetical. Algorithmic bias in health screening has been documented globally. Studies have shown that clinical risk prediction tools in the US underestimated health needs for Black patients because they used healthcare spending as a proxy for health needs - and spending was lower not because needs were lower, but because access was worse. The same structural dynamics exist in New Zealand for Pacific communities. We've written about this in the context of Māori health outcomes as well.
8.1%
of New Zealand's population identify as Pacific peoples, while Pacific representation in tech roles remains a fraction of this
Source: Stats NZ, 2023 Census
What Pacific Women Bring to AI
I get uncomfortable with "diversity makes better products" framing sometimes. It can reduce people to their demographic category, as though the point of having a Pacific woman on the team is to tick a box.
So let me be specific about what Pacific women with backgrounds like mine actually bring to AI work.
Community governance knowledge. I hold a matai title. I understand how collective decision-making works in practice, not theory. I've participated in village fono where decisions about resources, disputes, and community welfare are made through structured deliberation. That knowledge is directly relevant to AI governance design, where most frameworks are theoretical and untested.
Health system experience. Many Pacific women in Aotearoa work in health, as nurses, community health workers, programme coordinators, health navigators. They understand how health data is collected, where it breaks down, and what it misses. That operational knowledge is critical for building AI health tools that actually work for Pacific communities.
Relational thinking. Pacific worldviews are relational. Identity, health, and wellbeing are understood through connections - to family, community, land, and ancestors. AI systems are typically designed around individuals. A Pacific perspective challenges that assumption and opens up design possibilities that individual-centric frameworks miss entirely.
Practical bilingualism. Moving between languages and cultural contexts daily builds a translation skill that's directly applicable to AI work: taking complex technical concepts and making them meaningful to non-technical communities. And going the other direction, taking community knowledge and translating it into forms that technical teams can work with.
The Women I've Worked With
I want to name something that statistics miss. The Pacific women I've worked alongside across my career have been extraordinary, and largely uncredited.
The community health workers in Samoa who maintained data quality in rural clinics despite impossible conditions. The programme coordinators at Health NZ who kept complex multi-stakeholder programmes running through COVID. The grants coordinators at Foundation North who managed relationships with dozens of community organisations simultaneously.
These women were doing systems thinking, data management, stakeholder coordination, and programme evaluation - all skills that AI teams pay premium salaries for when they come packaged with a computer science degree and the right job title.
We don't need Pacific women to adapt to tech. We need tech to recognise the skills Pacific women already have. The knowledge of how communities work, how data connects to people, how decisions affect families - that's not soft skills. That's the hard part. The technology is the easy part.
Dr Tania Wolfgramm
Chief Research Officer
What Needs to Change
I'm not going to list five neat recommendations. The situation is more complicated than that.
But I will say this: the AI industry in Aotearoa is small enough that individual decisions matter. Who you hire. Who you listen to. Whose experience you value. Which communities you consult, and whether you actually change anything based on what they tell you.
At RIVER, I'm in a position to influence how AI systems are designed and evaluated. That matters. Not because I represent all Pacific women, but because I bring knowledge that changes what questions get asked. When I'm in the room, we ask different questions about data sovereignty, community impact, and cultural appropriateness. Not because I give a speech about it, but because those questions are part of how I think.
Every AI team in Aotearoa that works with health data, education data, social services data, or government data is making decisions that affect Pacific communities. If nobody on the team has lived experience of those communities, there are questions that simply won't get asked.
And in AI, the questions you don't ask are the ones that cause the most damage. If you want to build AI that works for all of Aotearoa, let's talk about how.