State of AI in New Zealand
Updated: 11 March 2026
Executive Summary
The past month suggests AI momentum in Aotearoa New Zealand is real, but uneven. The strongest signal is in healthcare delivery, where the Government says AI scribes are now live across every emergency department in the country, alongside a fresh move to explore AI-assisted breast screening. At the same time, AI governance is shifting from abstract debate to practical safeguards: the Privacy Commissioner has joined an international warning on AI-generated imagery, and Parliament’s online-harm inquiry has explicitly recommended bans on “nudify” apps and non-consensual deepfake sexual imagery. Outside government, the picture is more mixed: there is continued movement in enterprise infrastructure and sovereign data/AI capability, but startup and investment news was relatively quieter than the policy and public-service story. Overall, March 2026 looks like a month where New Zealand’s AI story broadened from strategy into frontline deployment, operational governance, and harm response. (beehive.govt.nz)
What Happened in the Past Month
Health became the clearest area of real-world deployment
The biggest concrete development was Health Minister Simeon Brown’s announcement on 28 February that AI scribe technology is now live in all emergency departments nationwide, with rollout completed to 1,250 ED doctors and frontline staff. The Government says pilot results showed doctors using the tool could see one additional patient per shift on average, making this one of the most tangible examples yet of AI moving from pilot to scaled public-service use in New Zealand. (beehive.govt.nz)
Just over two weeks earlier, on 12 February, the Government also said Health New Zealand is inviting proposals on AI image-reading for BreastScreen Aotearoa. That is not a deployment yet, but it is meaningful: it shows AI use in health is expanding from clerical support into possible clinical-assist workflows, with breast screening framed as a workforce and service-capacity issue rather than a tech experiment. (beehive.govt.nz)
Harms, privacy, and deepfakes moved further up the agenda
On 23 February, the Office of the Privacy Commissioner said it had co-signed a joint international statement warning about AI systems that generate realistic images and videos of identifiable people without their knowledge or consent, explicitly highlighting risks including non-consensual intimate imagery and harms to children and vulnerable groups. (privacy.org.nz)
That same concern is now showing up in Parliament. A parliamentary inquiry into online harm and young people, completed in early March, recommended that New Zealand ban “nudify” apps and prohibit the creation and distribution of non-consensual deepfake sexual imagery, while also calling for more New Zealand-based research on online harms. That does not amount to a full AI law, but it is a notable sign that AI policy attention is narrowing onto specific harms where the case for intervention is strengthening. (www3.parliament.nz)
Public-service AI machinery kept maturing behind the scenes
Public documentation on the New Zealand Public Service AI work programme to 2027 shows government agencies building more shared infrastructure for adoption, including plans for a central AI hub, an AI innovation and accelerator lab, an AI sandbox, a public-service assurance model, and a Govt.nz AI search assistant. The significance here is less about any one launch and more about the shape of the programme: central government is clearly trying to move from high-level principles toward reusable tools, common cases, and assurance mechanisms. (digital.govt.nz)
At the local-government level, Invercargill City Council’s Finance and Policy Committee in February recommended adoption of an AI Use Policy, with discussion covering staff AI literacy, reporting mechanisms, and data sovereignty. The accompanying guide also states that approved council AI must uphold New Zealand data sovereignty principles, including Māori data sovereignty. This is local and early-stage, but it matters as evidence that AI governance is diffusing beyond Wellington and into ordinary institutional policy-making. (icc.govt.nz)
Research and labour-market analysis remained active
A useful national research contribution came from the Reserve Bank of New Zealand on 3 February. Its analytical note used ten LLMs to estimate AI and robotics exposure across New Zealand occupations and found that AI exposure is highest in professional, managerial, and administrative work, while around 30% of workers face high combined exposure to AI and robotics. That is not a policy announcement, but it is one of the more substantive recent attempts to ground New Zealand’s AI debate in local labour-market evidence rather than imported assumptions. (rbnz.govt.nz)
Enterprise AI activity continued, especially around sovereignty and infrastructure
The clearest enterprise/infrastructure signal came from the HPE–2degrees announcement in late January, which remains relevant this month because it points to where larger New Zealand firms are heading: a private AI platform designed to keep critical customer and operational data onshore in New Zealand. The stated initial use cases include autonomous network operations, predictive maintenance, and AI-powered capacity planning. That is still a vendor-led announcement, but it fits a broader pattern in New Zealand’s AI market: adoption is increasingly tied to data residency, security, and infrastructure control, not just generic experimentation. (hpe.com)
National Direction and Near-Term Trend
The national line of travel is fairly clear: 2025 was about setting direction, while early 2026 is increasingly about operationalising that direction. The July 2025 national AI strategy set a pro-adoption, light-touch policy posture, and the January 2025 Public Service AI Framework established baseline guidance for agencies. The activity now showing up in health, public-service tooling, and SME support looks like the next phase of that agenda. (beehive.govt.nz)
That is especially visible in the research and capability pipeline. MBIE’s Artificial Intelligence Research Platform is now in phase two, with proposals due 31 March 2026, a platform funding decision scheduled for late April 2026, and investment decisions to follow in May 2026. In other words, last year’s headline announcement of a national AI research platform is now moving through the machinery of selection and funding rather than remaining purely rhetorical. (mbie.govt.nz)
There is also a quieter but important broadening in who AI policy is for. The Government’s AI Advisory Pilot for small businesses, announced in January, is designed to help SMEs with adoption barriers including privacy, data management, and responsible use. That complements what is happening in large agencies and enterprise IT: AI in New Zealand is no longer just a conversation for frontier researchers and big corporates. (beehive.govt.nz)
What Looks Quiet, Unchanged, or Early
For all the activity, this was not a month of major new national AI legislation, a new specialist AI regulator, or a blockbuster local startup funding wave. The strongest movement was in deployment and governance, not in legal overhaul or dramatic capital formation. That is important context: New Zealand’s AI story still looks more like adoption and integration than frontier-model development. (beehive.govt.nz)
Education and ecosystem activity remains active, but still leans more toward capability-building than system transformation. AI Forum NZ spent February preparing an AI Blueprint refresh, creativity summits, and an overseas mission focused on agentic AI, while Day of AI Aotearoa continues to push toward free foundational AI literacy for every school, teacher, and learner by the end of Term 1 2026. Those are positive signs for ecosystem maturity, but they are still upstream from nationwide institutional change. (aiforum.org.nz)
Overall Assessment
If February-to-early-March is the test, then AI in New Zealand is no longer best described as speculative. It is showing up in ED workflows, screening-service planning, public-service operating models, privacy enforcement, parliamentary harm debates, and telco infrastructure strategy. That said, the momentum is still concentrated in a few lanes: health is leading, public-sector governance is catching up, and enterprise adoption is increasingly shaped by sovereignty and security concerns. (beehive.govt.nz)
The simplest reading of the month is this: New Zealand’s AI ecosystem is broadening, but not yet exploding. The country is moving past the strategy-only phase and into practical implementation, while still lacking the scale of investment, infrastructure, and startup churn seen in larger markets. For now, that makes March 2026 look less like an AI boom and more like something more credible: measured institutional adoption, with sharper attention to public trust and harms. (beehive.govt.nz)