Building AI Capability for National Transformation in New Zealand

AI is no longer a future conversation in New Zealand. It is becoming the operating system for how work, government, institutions, and services are designed.

But the real shift is not technological. It is structural.

Across industries, from call centers to regulation, from public services to national productivity, a clear pattern is emerging. AI is not simply replacing people. It is reshaping how work is organized, how decisions are made, and how outcomes are delivered. The organizations that succeed will not be those that merely adopt AI tools, but those that rebuild capability around people, process, and systems working together.

At the centre of this shift is a critical question. How do you build national AI capability in a way that improves productivity, strengthens public services, and keeps people at the centre of the system?

Because the transformation is already underway, the only question is whether we design it or react to it.

Inside the Digital Sphere, our guest is Madeline Newman, Executive Director of the AI Forum New Zealand. She brings extensive global experience across AI, regulation, mental health, and financial services innovation in the UK, including a key role in shaping the internationally recognized RegTech ecosystem at the UK’s Financial Conduct Authority.

Today she leads the organization driving the delivery of New Zealand’s strategic response to AI aligned to the vision set out in the AI Forum’s AI Blueprint for Aotearoa. She also serves on a number of AI advisory boards and mentors emerging leaders through programs like the University of Auckland’s Kupe Leadership program. She is focused on steering New Zealand toward a more innovative, responsible and inclusive AI-capable future.

This conversation explored AI adoption and national strategy in New Zealand, highlighting that AI is fundamentally a people and organizational change challenge rather than just a technology shift, and how organizations can use AI to solve real business problems rather than adopt technology for its own sake. It discussed the role of regulatory sandboxes in enabling safe experimentation and how AI can improve productivity through augmentation, process simplification, and reinvestment of gains. Overall, the key message was that successful AI adoption depends on building trust, aligning regulation with innovation, and ensuring real economic and social impact.

You’ve spent over two decades working in environments where technology directly impacts people’s lives - from regulated industries to digital mental health. How has that shaped your belief about how AI should be designed and deployed in real organizations, especially when it comes to things like customer-facing systems?

In my work in operational architecture, I have always come back to one core idea. At its heart, it is really about people, process, and systems working together. Humans are always the core component of any system, and understanding how they function together is essential. When you look at it from a societal or sociological perspective, it becomes clear that it is about how people operate within their cultural and organizational norms. When you try to change how people work, you are effectively shifting those norms, which takes time, effort, and careful consideration. It is rarely just a technology change. It is a fundamental shift in Behavior and mindset.

Over the past 20-plus years, I’ve seen a major shift with AI bringing us closer to a model where the people doing the actual work are also becoming part of designing the systems they use. A good example of this is call centers, which are often the front door of an organization and the first point of contact for customers. Traditionally, performance in call centers has been measured in very operational terms - how many calls are answered, how quickly each call is handled, and how long customers are waiting in the queue. The entire system is designed around speed and throughput, with constant pressure on call centre workers to resolve issues quickly and move on to the next call.

Now, when you introduce an AI assistant into that environment, you can fundamentally change how that work is done. You take all of the company’s documentation, processes, procedures, and knowledge base, and embed it into an AI system that supports the call centre worker in real time, helping them answer customer queries and guide conversations more effectively. However, the important part is that you don’t build this system in isolation - you involve the call centre workers themselves in the design process. These are the people who interact with customers every day, and they understand the real problems, the nuances, and the edge cases better than anyone else.

When you do this, two important things happen. First, the quality of the system improves significantly because it is grounded in real-world experience and practical needs. Second, adoption becomes much smoother because the people using the system have already been part of building it, so they are already invested in its success and understand its value.

Once this kind of system is in place, you often see traditional performance metrics start to break down. For example, instead of focusing on how many calls are answered or how quickly a call is closed, the first person who picks up the call can often resolve the entire issue end-to-end. They don’t need to transfer the customer or rush the conversation. While calls take longer, the customer doesn’t need to call back multiple times, and their issue is fully resolved the first time.

As a result, the old metrics like call volume, speed, and queue time are no longer the best indicators of success. Instead, customer satisfaction and first-contact resolution become far more important. This represents a shift from efficiency-only thinking to outcome-based thinking.

Ultimately, what AI enables here is not just automation, but augmentation. It enhances the capability of the worker by giving them immediate access to the organization’s knowledge, embedded directly into their workflow. And when you step back, you realize this is not purely a technology transformation. It is fundamentally a business and people transformation, where technology acts as the enabler for a new way of designing and delivering work.

From your experience with the UK Financial Conduct Authority’s RegTech work, what role do regulatory sandboxes and tech sprints play in helping organizations test new technologies safely, improve consumer access, and reduce the gap between regulation and innovation?

The sandbox concept is a really valuable one, and I’m pleased to see that the Financial Markets Authority (FMA) is actively running a sandbox environment. What a sandbox allows is for startups and established organizations alike to trial new technologies or innovative ways of working that may not have been used before. While we can do a great deal of testing in controlled environments, a sandbox provides the opportunity to test solutions in the real world, under real regulations, with regulators working alongside you. That creates a unique learning environment where both the organization and the regulator can observe outcomes together. It becomes particularly useful when you're operating on the edge of regulation - where it’s not entirely clear what is acceptable and what isn't. Those situations often reveal opportunities to refine or improve regulatory frameworks.

We saw this firsthand in the UK through the work of the Financial Conduct Authority (FCA). Around 2015 and 2016, we ran a series of technology sprints, essentially hackathons focused on regulatory innovation. At the time, one of the major challenges facing banks was the enormous cost of maintaining legacy technology systems. In many cases, more than 90% of technology change budgets were being spent simply keeping existing systems operational.

That left very little available for customer-focused innovation. What little remained was often consumed by regulatory compliance requirements, especially in the years following the Global Financial Crisis. The UK, which was still part of the European Union at the time, experienced what we referred to as "peak regulation," where a significant wave of new compliance requirements emerged. Many organizations found themselves shifting an additional 10% to 15% of their budgets toward regulatory compliance. We were interested in regulatory technology, or RegTech, as a way to reduce compliance costs while still meeting obligations.

As regulators, we couldn't tell organizations exactly what technology to use, nor could we recommend specific solutions. What we could do was create environments that encouraged innovation and helped organizations explore possibilities. That's exactly what these technology sprints were designed to achieve.

One of the first sprints focused on digital access. At the time, people without access to digital financial services were often missing out on savings of around £400 to £500 per year. For that call centre worker earning £12-16,000 annually, which was close to a living wage in the UK at the time, that represented a significant financial disadvantage. We wanted to explore how banks could use technology to improve digital access and inclusion.

During that process, we invited a company called iProov, which had developed a security- focused facial recognition capability. Today, facial recognition may seem commonplace, but at the time, it was a groundbreaking idea. What made this particularly interesting was the reaction from the major banks. Their compliance teams were surprised that a regulator would actively invite innovative technology providers into discussions about improving customer access.

The technology sprint itself was successful, with six of the proofs of concept going on to further development, but what happened afterwards was perhaps even more important. Within a couple of months, many banks began expanding their use of biometric authentication, including fingerprint identification on mobile devices. What became clear was that the barriers were not necessarily regulatory. In many cases, organizations had interpreted the regulations more conservatively than required. The message we had sent was that the regulator wasn't preventing innovation - internal interpretations of the rules were.

Another area where sandbox environments have proven valuable is robo-advice. Many people with smaller investments simply cannot afford the services of an independent financial adviser. Yet having some form of guidance is often far better than having none at all. Digital advice platforms can provide accessible, cost-effective support to a much wider group of people. The challenge is ensuring that this advice is delivered safely, responsibly, and within appropriate regulatory boundaries. Sandbox environments allow organizations to experiment with these solutions, test them under regulatory supervision, and build confidence that they can deliver value to customers while maintaining appropriate protections. Ultimately, that's one of the greatest strengths of a sandbox - it creates a safe space where innovation and regulation can evolve together.

You’ve worked across both agile, early-stage startups and large, established institutions. When a traditional organization wants to deploy emerging technologies like AI, what is typically the biggest cultural hurdle they face, and how do you help them overcome it?

The biggest hurdle in all of this is actually the people change. It’s getting people to adopt new ways of working and then making sure it actually sticks. That’s always been the hardest part of change management. You can introduce tools, you can introduce systems, but getting people to genuinely shift how they work - that takes time.

And I think one of the things people underestimate is just how long that cultural shift actually takes. When we talk about cultural norms, or the “unwritten rules” of how people behave at work, those don’t change overnight. In most cases, you’re looking at around 18 months at a minimum for a team to really settle into a new way of working.

So the tools you design, especially things like AI systems, need to support that journey. They need to help people make that transition, not just assume it’s going to happen automatically.

One of the mistakes I see quite often when organizations implement AI or similar toolsets is that they focus too much on the technology and not enough on what happens after it’s introduced. Xero completed a survey in April that showed something quite interesting. In the areas where adoption was high, teams were seeing around a 20% time saving in an average week.

Now, that sounds great on paper. But here’s the issue - it doesn’t automatically translate into full role savings. So if you’ve got a team of 10 people, it doesn’t suddenly mean you can remove two people or reclaim that time cleanly.

What often gets missed is a proper benefits realization plan that accurately tracks where the time is being saved so that you can consciously decide what replaces that effort. If you don’t, people will naturally just move on to the next most important task. But that “next most important task” might not be what you intended them to focus on, meaning you end up with an illusion of efficiency - where the time is saved on paper, but in reality, it just gets absorbed elsewhere in the system.

And that’s why the discipline around change management is so important. It’s not just about introducing AI or new tools. It’s about deliberately designing what work gets removed, what gets replaced, and how you make sure those gains actually stick in a meaningful way.

The International Monetary Fund highlights that New Zealand’s GDP per hour worked has historically trailed peers such as Denmark and Finland. How do you think AI can help reshape New Zealand’s long-standing productivity challenge?

The Treasury has started talking about this quite openly. If you look at the trend, productivity gains have been on a downward path since the 1970s. There has been some growth, but when you compare New Zealand to a country like Denmark, the difference is quite striking.

In the 1960s we were actually quite similar - both relatively small, rural-based economies, with similar population sizes - but today Denmark is a world leader in productivity, and New Zealand simply isn’t at that same level.

So the question is, what do we actually do about that? I think AI presents a two-pronged opportunity to reshape that productivity challenge.

The first is pretty straightforward - using AI to augment the workforce we already have. Radiology is a good example. We don’t have enough radiologists to meet demand, but we can use AI tools to support and augment the radiologists we do have, helping them interpret scans faster and make better decisions. That’s a very practical example of where AI can genuinely move the dial in terms of capacity and efficiency.

But there’s a second part to this that often gets missed.

In businesses, you don’t actually see productivity growth just from automation alone. When we automate processes - whether that’s using AI or any other form of technology - and reduce headcount but then don’t reinvest those savings back into the business, nothing really changes at a system level.

If people are made redundant, graduates aren’t hired, and the savings made are not reinvested, then you don’t actually see growth; you might see the same output with fewer people. From a broader economic perspective, that can actually look like contraction rather than expansion. Reinvestment is critical if you want to shift the dial on GDP growth and productivity. Those efficiency gains have to be consciously reinvested into growth - whether that’s new skills and capability, new products and services, or new people.

AI is an enabler for productivity gains, but the decisions on whether and how to invest savings generated in growth are leadership decisions.

You have seen that years ago, deploying AI required a million-dollar custom setup, whereas now small local businesses can adopt it for a fraction of that cost. How has this dramatic ‘democratization’ shifted the tech conversation across New Zealand organizations?

I joined the AI Forum in March 2022, before the free ChatGPT release of November, which changed everything. AI was producing real results, a lot of interesting work was happening in machine learning, and we were seeing a small number of large-scale single deployments of things like digital humans. But the investment required was massive, millions of dollars per deployment, which basically meant only large corporates could play in that space.

Then generative AI came along and completely shifted that dynamic. It made everything far more accessible and dramatically sped up the test and release cycles. Suddenly, within a short amount of time, we had low-code, no-code - or vibe coding, where people can just think through an idea and start building. Now, almost anyone can sit down and actually create something that uses AI.

To give you a sense of how things have changed, in 2021 I was Head of Innovation and Product for a wellbeing platform in the UK, and we had a really strong team of developers based in Ukraine. They were incredible - they could deliver a new product in about six weeks, which was extremely agile at the time. But now, with these tools, I can prototype something in two hours.

We saw this in practice recently at an AI Solutioneering event we ran at the University of Auckland’s Unleash Space, in collaboration with the Auckland Tech Council, Auckland Business Chamber and Deloitte during Tech Week. We asked 100 companies to bring their own problem to solve, had them build their own solutions in 2 hours using AI tools, supported by technical experts and mentors.

While this didn’t magically solve their deep challenges in a single session, it did build confidence and familiarity. It showed business owners and leaders that AI isn’t this complex, unreachable thing. It’s something they can actually start using right now.

And that’s important, because it opens the door for companies like Voxitec to step in and say, “Okay, let’s sit down and really understand your business problems and figure out how AI can solve them properly.”

The key lesson in all of this is that you should always start with the business problem, not the technology.

An interesting thing we’re seeing in New Zealand right now is a low-trust, high-use environment. People are using AI, but they don’t fully trust it yet. And part of why we ran that AI Solutioneering event was to start building trust.

It’s similar to social media. People don’t necessarily trust social platforms, but they still use them every day.

Low trust means that people are not ready to embed AI deeply into their core business processes. People are experimenting with it and seeing some benefits as a result, but the examples of real value we see are where organizations have built it into how they operate - back to people, processes, and technology.

There are quick wins available early - such as using generative AI instead of manually writing notes or recording and reviewing meetings. But the real shift happens when AI is embedded into core business processes and tailored to the organization, and that still requires some investment - and trust.

The AI Forum has been working closely on outlining key action areas for a national AI blueprint. What are the most critical pillars of that vision that New Zealand must get right over the next few years?

The AI Blueprint for Aotearoa is the AI Forum’s idea of what the nation’s AI Strategy could be, and we deliver actions against that strategy. We focus on 5 Pillars and deliver to those through a number of work-streams, mainly sector-focused, with two newer work-streams that reflect current conversations and concerns - Sustainable AI and Social License. Both of which have become increasingly important recently as AI adoption continues to grow and organizations start thinking more seriously about long-term impacts and governance.

The AI Forum's vision is of leading AI for a thriving Aotearoa that is innovative, responsible, and inclusive. Sustainability is naturally embedded within the idea of responsible AI.

The first pillar is about understanding new and evolving capabilities, exposing new opportunities, and supporting responsible adoption. The work in this area is about helping people discover opportunities, understand what's possible, and adopt AI in a way that delivers value while managing risk. That means encouraging experimentation, supporting innovation, and helping organizations prototype new capabilities.

We've done things like sandboxing initiatives and developed what we call a ‘living white paper’. Instead of being a static report that becomes outdated over time, it has trusted sources it can use to keep up to date, and it's something people can interact with. You can ask questions and generate the information you need, effectively creating a version of the white paper that's tailored to your specific interests.

Alongside that, we're also supporting responsible development through governance frameworks, policy guidance, and practical approaches to data and AI management.

The second pillar is about increasing capability and scaling innovation. Helping organizations to adopt AI responsibly at scale. That means providing access to tools, services, infrastructure, and the support systems needed to make innovation sustainable.

There's a lot of discussion at the moment around sovereign AI and what it means for New Zealand to have local capability and control over critical AI infrastructure. These conversations are becoming increasingly important because they influence what opportunities are available to New Zealand businesses in the future.

We're also continuing to build on the work we've already done. Back in 2023, for example, we launched our AI Governance website because there simply wasn't enough New Zealand-specific guidance available.

By 2025, more than 200 member volunteers were creating content that we make freely available to help organizations adopt AI responsibly while continuing to innovate. Beyond that, we're also looking at research and development, attracting investment, and creating programs that help New Zealand businesses grow and generate long-term value.

The third pillar focuses on adoption and risk management, which comes down to trust, social license and access. One of the biggest challenges we face is ensuring AI doesn't deepen existing inequalities.

The technology has enormous potential, but if we're not careful, it could also widen digital divides and create new barriers for people who are already disadvantaged. That's why inclusion is such an important part of our work. We need to make sure the benefits of AI are shared broadly across society and that we're bringing everyone on the journey, not just a small group of organizations or early adopters.

The fourth pillar is about talent. One of the most important things we can do is ensure people have the skills they need to participate in an AI-enabled economy. That means preparing future generations and making sure students leave school with a strong understanding of AI and digital literacy.

Equally important is supporting the workforce we already have. Existing employees need opportunities to up-skill, re-skill, and adapt as jobs evolve. If we want everyone to share in the rewards that AI promises, we need to invest in people and give them the tools and capabilities to succeed.

New Zealand already has an exceptional talent pool. In fact, companies are choosing to establish operations here because of the quality of our workforce. There is a scarcity of AI talent worldwide, so the challenge is not only building talent but retaining it by continuing to create opportunities for growth and development.

The fifth pillar is global reach, leading with our strengths to build international connectedness and develop new markets. This is about ensuring New Zealand remains connected to international markets and can compete successfully on a global stage. It includes helping organizations understand regulatory requirements across different jurisdictions and making it easier for exporters to enter those markets.

For example, if you design your systems to be GDPR compliant from the outset, you've already met many of the requirements needed to operate across Europe and a number of other regions.

Global reach is also about showcasing New Zealand's unique strengths. Indigenous AI is a particularly good example. Around the world, there are very few serious conversations about Indigenous approaches to AI that don't involve Māori voices.

Pacific communities are also increasingly contributing to this discussion through initiatives such as Fibre Fale and our own Pacific Advisory Forum. As Frances Valentine often points out, the workforce of the next 10 to 20 years will look very different from the workforce we have today, with Māori and Pasifika communities representing a much larger proportion of New Zealand's workforce. That's why it's so important that everyone is included in this journey.

Global reach also extends to international collaboration and standards development. New Zealand has been actively involved in helping shape international AI standards rather than simply adopting standards developed elsewhere.

One example is the work undertaken by New Zealand's AI standards committee, which worked with their counterparts in Australia to ensure international standards reflected the needs and perspectives of both countries. As a result, in 2025 New Zealand and Australia became the first countries in the world to adopt the full suite of those ISO AI standards.

We've also contributed to the International AI Safety Report, which provides a scientific foundation for global conversations around AI safety. The AI Forum is one of 100 organizations from around the world that contribute to and review that work. So while it can sometimes feel as though nothing is happening, there is work taking place to provide evidence-based commentary on the potential threats posed by AI, helping countries to develop policy, guidance and guardrails to ensure AI is developed in a way that is safe, responsible, and beneficial for society.

When we look at global AI governance models, some lean heavily toward strict regulation, while others favor open market innovation. What kind of balanced governance model is right for Aotearoa?

It's a difficult question to answer because we're not a political organization and we don't set policy. Our role is really to provide advice, guidance, and recommendations. The regulatory approach taken in New Zealand was essentially not to introduce any new AI-specific regulation at the time. That doesn't mean AI isn't regulated - it simply means that the government chose to rely on existing laws and frameworks rather than creating entirely new ones. From the AI Forum's perspective, that was a reasonable approach at the time.

What's interesting, though, is that the conversation has evolved quite significantly. We know a lot more about AI today than we did a few years ago. We've seen how these technologies are being used, we've learned more about the risks and opportunities, and we've developed a much better understanding of the different ways regulation could be approached.

One person who talks about this particularly well is Professor Alexandra Andhov from the University of Auckland, where she serves as Chair of Technology and Law. She often describes regulation as a verb rather than a noun. I think that's a really powerful way of looking at it because one of the biggest challenges with regulating AI is that the technology moves incredibly quickly. By the time you've fully defined something, it's already changed. That's one of the reasons our government has been cautious about introducing new regulations too early. They didn’t want to accidentally stifle innovation or lock themselves into frameworks that quickly became outdated.

We can see this debate playing out internationally as well. If you look at the European Union's AI Act, supporters argue it's necessary to create safeguards and accountability, while critics worry it could slow innovation. Neither side is entirely wrong. It's a difficult balance to strike.

The New Zealand government’s stated preference was to be a "fast follower", taking advantage of observing what happens elsewhere before making any decisions. While definitely not without risk, this approach can provide the opportunity to learn from both the successes and the mistakes of other jurisdictions. That's valuable because regulation can be a very blunt instrument. Once legislation is introduced, it can be incredibly difficult to unwind or adjust.

That's why I find Alexandra's idea of regulation as a verb so compelling. Regulation shouldn't be viewed as something static that gets written once and then sits on a shelf. It needs to evolve alongside society, technology, and changing expectations.

What's particularly interesting is that AI itself may help us rethink how regulation works. Imagine being able to model a regulatory environment and run thousands of scenarios before implementing a change. You could ask questions like, "What happens if we adjust this regulation?" or "What would the likely outcome be if we introduced this requirement?" You could simulate different impacts across industries, communities, and society more broadly.

The potential is to eventually be able to start asking a different kind of question. Instead of asking, "What happens if I make this change?" you could ask, "I want this outcome - what changes are most likely to help me achieve it?"

Take something like social media restrictions for children under 16. An age limit is one possible approach, but it's a fairly blunt tool. It might improve outcomes to some extent, but it's unlikely to achieve perfect compliance. So the question becomes, are there smarter ways to achieve the same objective?

I don't think we have all the answers yet, but I do think we have an opportunity to start thinking about regulation differently. In business, rather than relying on retrospective compliance - where someone checks what you've done after the fact and penalizes you if you've made a mistake - we should be moving towards more proactive and embedded forms of compliance, where controls are part of the process, and the results become part of your customers’ quality expectations.

Instead of regulation being something that sits outside a process, regulation can become part of the process itself. Built in from the beginning, providing guidance and guardrails as decisions are being made. That's one of the places where I think some of the most interesting opportunities lie for the future of both AI and regulation.

Educational institutions are evolving rapidly. How must our tertiary education pathways shift to become more agile and responsive to industry demands for AI and data infrastructure roles?

There’s a big conversation happening about this at the moment, and I often point people toward Professor Albert Bifet. He heads up the AI Institute at the University of Waikato and is also a professor at Telecom Paris. What’s interesting about the Telecom Paris setup is their strong link to industry, delivering an industry-facing university model where the focus is very much on developing skills and capabilities needed in the real world.

Courses are continuously reviewed and updated, so instead of designing curriculum in isolation and then leaving it unchanged for years, they’re constantly adjusting it based on what the industry is actually asking for. It’s not about what academics think might be relevant in theory - it’s about the skills people actually need in the market right now.

And I think that approach is really important, especially when you think about how quickly things are changing with AI. In some professions today, generative AI tools are already becoming standard. So if graduates are coming through education systems without exposure to those tools, then there’s a gap forming. And people are starting to recognize that.

There are significant challenges in rewriting qualifications, assessments, and curricula at pace. Our education systems aren’t designed to change that quickly. But there is an emerging willingness to engage in the conversation, and thought leaders in this field are already exploring how some qualifications can evolve to reflect more dynamic, real-world skill requirements.

We are seeing staggering statistics where some ‘AI-intensive’ companies spend thousands per month on AI infrastructure per employee. What are your thoughts on this macroeconomic shift, where capital is moving away from human payroll and into hyperscaler cloud computing and AI infrastructure?

There is certainly an ongoing debate about investment in people vs machines, but (large LLM organizations aside) for now anyway the answer lies in AI augmentation, not replacement of people.

We recently sent a delegation to China, and when they came back, one of the key observations they shared was quite interesting. In some cases, it’s illegal to replace people with a robot or AI. The idea is that AI is being introduced, and productivity is being enhanced, but jobs are not being cut as a direct consequence of it. Delivering a model where AI is being implemented very aggressively, but it’s being done alongside a strong protection of employment.

As I said, it’s not really an “AI versus people” conversation. It’s an AI and people conversation. Growth, investment, and productivity only really work when those two are aligned, not when one replaces the other in isolation.

Right now, we’re also in a very unusual pricing window for generative AI. It’s extremely cheap compared to what it actually costs to produce - and that points to potential price increases in the future. In fact, you could argue it’s never going to be this “dumb” or this inexpensive again. The current indications suggest it’s going to become more capable, but also more expensive over time.

Accelerating the digitization of the public sector is a massive opportunity to reduce costs and enhance public services. Where do you think the government should start its internal AI transformation?

Brandon Hutchison has put together a really interesting piece of work - something like 160 different suggestions around how we could improve this space 0 and it’s gaining quite a lot of traction. It’s attractive because the ideas are generally very practical, very grounded, and he’s also gone as far as estimating the potential cost and impact, which is really important in this kind of discussion.

But the key point here isn’t about replacing people. It’s about helping people do a better job and, just as importantly, removing a lot of the unnecessary complexity that builds up in systems over time.

A good example would be a person with a long-term irreversible condition having to regularly reapply for a disability benefit, including providing a doctor’s certificate, wasting health resources as well.  It’s not just a small inconvenience. It’s the doctor’s time, the patient’s time, travel, admin, paperwork - all of that effort just to confirm something that hasn’t changed. Couldn’t we connect those dots in a smarter way?

AI is really good at solving riddles and seeing patterns - untangling the systems we’ve created over time and ensuring that they deliver our expected outcomes in a way that makes things simpler, more efficient, and less burdensome for everyone involved.

Arthur C. Clarke, in his 1962 book Profiles of the Future, reflected on the idea that technology could increasingly extend and transcend biological limitations. Looking ahead to the next decade, do you see humans and AI remaining distinct collaborators, or moving toward a more deeply integrated, symbiotic relationship?

I have to say I was a big Arthur C. Clarke and Isaac Asimov fan growing up!

Professor Michael Witbrock from the University of Auckland is a leading authority in brain-computer interfaces, and when I’m sitting there using my phone, he’ll literally say, “Put that down. In a few years, you’re not even going to need it. You’ll just be able to think it, and it will happen.”

And the thing is, he’s not saying that as science fiction. He’s actually working on that future.

You can already see early versions of it in things like advanced prosthetics and limb replacements. That kind of “Luke Skywalker moment” where a prosthetic hand responds, opens, adjusts, and feels like an extension of the person - that’s not fantasy anymore. That’s in development right now.

More broadly, it raises profound questions about where the boundary lies between thought, intent, and action. How comfortable are people with the idea of systems that can interpret what’s happening inside their mind? And how does that balance change when there is an obvious benefit, like the restoration of hearing? These are real questions for our near future.

There are incredible technologies that restore capability and can improve quality of life in very real ways. These very practical assistive technologies that are already improving people’s daily lives are also starting to blur the line between thought and action. Right now, we’re somewhere in the middle of that transition - with a lot to think about!