AI first caught my attention because I could see its potential to save people time, reduce overwhelm, and help them navigate information more easily.
I'm all about efficiency and creating a clear flow that is easy to use and guides people towards what they need. The opportunity to bring those things together in a way that could empower people was what interested me most.
As Tai Ora grew beyond its original product tools, I envisioned how AI could help people navigate information, recognise patterns, and find pathways that might support them.
But wellbeing is personal, and that made me question what would happen if the technology misunderstood someone, gave the wrong guidance, or spoke with confidence about something it didn’t really know.
After questioning how much of the information online we could really trust, the last thing I wanted was to build another system people were expected to trust simply because it sounded certain.
I didn’t want Tai Ora’s AI to pretend it had every answer.
I didn’t want it diagnosing people, making decisions for them, or directing them down a pathway simply because a system had decided that was where they belonged.
The more personal the information could become, the more important those boundaries felt.
iGlo™ can contain photos, notes, moods and details about someone’s experience over time. Future Tai Ora pathways could involve confidence, identity, finances, whakapapa and other parts of life that people may not feel comfortable sharing anywhere else.
If people were going to trust Tai Ora with those parts of themselves, that trust needed to be protected from the beginning.
That meant privacy couldn’t be something added later. People needed to know that their information belonged to them, that their private experiences would remain private unless they chose otherwise, and that their data wouldn’t simply become something to sell.
But protecting people’s privacy was only one part of it. I also had to recognise the limits of the technology itself.
AI can produce an answer that sounds convincing even when it is incomplete or wrong. In beauty and wellbeing, that could leave someone following advice that wasn’t right for them. In more sensitive areas of life, the consequences could be much greater.
I was thinking about how AI could support Tai Ora without taking over someone's experience.
Its role would sit between the human-led lessons, videos and wellbeing journeys, helping people reflect on what they had learnt and guiding them towards a possible next step.
The AI wouldn’t create the pathways itself or decide where someone belonged. It would help people navigate pathways that had already been carefully developed and supported by real people.
There also needed to be times when the AI stopped.
If something required medical, legal, financial or emergency support, the right response wasn’t to keep generating advice. It was to be clear about its limits and guide the person towards the right kind of human help.
Human oversight mattered too. Technology can help organise information and recognise patterns, but people still need to be involved in deciding how the system develops, reviewing what it produces, and responding when something doesn’t work as intended.
I’m also aware that AI has an environmental cost. The data centres behind it require significant energy and computing power, and many also use water for cooling.
Tai Ora isn’t designed to use AI for every function or interaction. Much of the experience will come through human-led lessons, videos and structured wellbeing journeys. AI will be used purposefully between those experiences, where it can help someone reflect, reduce overwhelm or find a possible next step.
If a simpler tool can achieve the same outcome, then AI may not be necessary.
For me, ethical technology isn’t about making AI sound friendly or adding the word “responsible” to a website. It is about the choices made behind it.
What information does it collect? Who controls that information? What is it allowed to suggest? What happens when it is unsure? And who is accountable when something goes wrong?
I want Tai Ora’s AI to support people without taking control away from them. It should help us understand our options, ask better questions and find clearer pathways, while leaving the final decisions in our hands.
Building an ethical AI ecosystem doesn’t mean claiming the technology will always be perfect. It means being honest about what it can and cannot do, creating boundaries before they are needed, and continuing to question whether the system is serving the people it was built for.
Because the most powerful technology isn’t the technology that makes every decision for us. It is the technology that helps us understand our options, ask better questions and make better decisions for ourselves.
