AI Adoption in Healthcare in New Zealand: A Living Whitepaper
Updated: 9 March 2026
Executive Summary
New Zealand healthcare is moving from AI exploration to selective operational deployment. As of March 2026, the clearest shift is that AI is no longer confined to small pilots: it is now live in every public emergency department via ambient documentation tools, while other use cases such as diabetes retinal screening, breast screening image support, and mental health navigation are progressing through pilots, procurement, or early implementation. At the same time, the Ministry of Health still describes the country as being in the early stages of assessing how AI can be safely adopted, which accurately captures the overall picture: adoption is real, but still narrow, supervised, and infrastructure-dependent. (beehive.govt.nz)
The pattern is now clear. In New Zealand, AI adoption is strongest where it can reduce clinician admin burden, speed repetitive interpretation tasks, or extend scarce specialist capacity without removing clinician accountability. The country is also building a thicker governance layer around that adoption, through Health NZ’s National Artificial Intelligence and Algorithm Expert Advisory Group, privacy controls that keep AI use in a closed Health NZ environment, and professional guidance under development from the Medical Council of New Zealand. (info.health.nz)
What Has Changed Since the Previous Edition
- AI scribes have moved from pilot to national frontline use. On 28 February 2026, the Government announced that AI scribe technology was live in all emergency departments nationwide, reaching about 1,250 ED doctors and frontline staff, with Health NZ also progressing approval of more than 1,000 additional licences, mainly for mental health teams. (beehive.govt.nz)
- Breast screening has moved into formal AI exploration. On 12 February 2026, Health NZ began an exploratory process for AI image-reading support within BreastScreen Aotearoa, explicitly linking this to service sustainability, workload pressure, and the screening age extension to 74. (beehive.govt.nz)
- The health system now has a digital modernisation frame for AI. In November 2025, the Government released the first Health Digital Investment Plan, established a Centre for Digital Modernisation of Health, and said Health NZ had created HealthX to accelerate innovation and AI use, including AI scribes, remote monitoring, and augmented x-ray processes. (beehive.govt.nz)
- Research funding has moved from intent to portfolio. The HRC’s 2025 AI in Healthcare initiative has now translated into 10 funded studies worth $4.6 million, creating a more visible pipeline from policy ambition to implementable use cases. (hrc.govt.nz)
- Primary care adoption has become more commercial and integrated. Survey evidence and product launches show AI scribes spreading in general practice, while Medtech now offers an integrated AI layer inside Medtech Evolution rather than a standalone add-on model. (pubmed.ncbi.nlm.nih.gov)
Current State of AI Adoption
1. Frontline Operational Use
Emergency departments: the most mature public-sector deployment
The most significant deployment in New Zealand healthcare is the nationwide ED scribe rollout. According to the 28 February 2026 Government announcement, doctors in the initial pilot were able to see one additional patient per shift on average because of time saved on documentation. Early feedback from Middlemore also found 80% of surveyed staff said the tool improved productivity or efficiency, and 84% said it improved their overall shift experience and wellbeing. (beehive.govt.nz)
Why this matters
- It is the first example of AI being deployed at true national scale in frontline public care in New Zealand. (beehive.govt.nz)
- It confirms that NZ adoption is currently strongest in documentation automation, not autonomous clinical decision-making. (beehive.govt.nz)
- It creates a template for adjacent rollouts, especially into mental health services. (beehive.govt.nz)
Diabetes retinal screening: AI as a capacity extender
In February 2025, Health NZ launched an AI-enabled diabetes retinal screening pilot in Māngere. The model uses trained community workers to capture retinal images, with AI performing the primary grading and ophthalmologists providing backup review during the pilot. The stated goal is to generate real-time screening results, increase screening volumes, shorten waits, and free specialists to focus on higher-value clinical work. (beehive.govt.nz)
Why this matters
- It shows AI being used to shift work closer to community settings. (beehive.govt.nz)
- It addresses a concrete backlog problem: the Government said more than 26,000 people in South Auckland had missed recommended screening in the previous two years. (beehive.govt.nz)
- It is one of NZ’s clearest examples of AI being used for triage and throughput, not just paperwork. (beehive.govt.nz)
2. Near-Term Clinical Expansion Areas
BreastScreen Aotearoa: formal validation phase underway
As of 12 February 2026, Health NZ is inviting organisations with AI image-reading experience to show how the technology could be used safely and effectively in BreastScreen Aotearoa. The release frames this as the first step in a validation process, not a deployment decision. It also links the work to demand growth, radiologist workload, the extension of screening to age 74, and the fact that roughly 270,000 women aged 45 to 69 are screened annually. (beehive.govt.nz)
Implication
- Breast imaging is likely to be New Zealand’s next major public-sector clinical AI domain, but only after validation, workforce engagement, and safety review. (beehive.govt.nz)
Radiology and acute stroke: strong implementation pipeline
The HRC-funded project “Transforming radiology in New Zealand - From accuracy to implementation of AI” will evaluate whether AI can improve the speed and accuracy of chest x-ray interpretation and reporting, while also studying workflow impacts, outsourcing costs, and implementation barriers. A second HRC project, “Optimising acute stroke care with artificial intelligence,” will measure the impact of AI-assisted rapid brain-scan interpretation on treatment rates and delays across NZ hospitals. (hrc.govt.nz)
Implication
- Imaging remains the strongest candidate for the next wave of clinical AI in New Zealand because it fits the country’s adoption logic: high-volume, specialist-constrained, measurable, and auditable. (hrc.govt.nz)
Primary Care and Private-Sector Adoption
AI scribes are spreading faster in primary care than in hospital specialties
The best current snapshot comes from a 2025 PubMed-indexed survey of NZ primary care providers. Among 197 respondents, 40% had experience with AI scribes. Reported benefits included less multitasking, time saved, and improved patient rapport; the most common concerns were compliance with NZ legal and ethical frameworks, data security, errors or omissions, and data leaving New Zealand. Only 59% of users reported seeking patient consent, and just 66% had read the tool’s terms and conditions. (pubmed.ncbi.nlm.nih.gov)
This survey is important because it shows that primary care adoption is not waiting for perfect policy clarity. Uptake is happening first, with governance, consent practice, and risk understanding catching up unevenly across providers. (pubmed.ncbi.nlm.nih.gov)
Medtech’s integrated model signals market maturation
Medtech Global now markets Medtech AI as an “AI intelligence layer” built into Medtech Evolution. The company says it can save six to eight minutes per consultation, generate structured notes directly into the PMS, support referral letters and correspondence, and keep the clinician in review-and-approve control. It also highlights “no audio retention” and privacy-by-design positioning. (medtechglobal.com)
What this signals
- NZ primary care is moving from generic transcription tools to workflow-embedded AI. (medtechglobal.com)
- Vendors are increasingly competing on integration, safety posture, and NZ-specific workflow fit, not just transcription accuracy. (medtechglobal.com)
- The dominant commercial use case remains documentation and communication augmentation. (medtechglobal.com)
Mental Health and Community Care
Mental health is emerging as a second major adoption lane after documentation. In November 2025, the Government funded Whakarongorau to develop a mental health AI navigation platform intended to act as a digital front door, helping people identify available support, understand local options, and in some cases book directly. The policy rationale is access and system navigation rather than diagnosis. (beehive.govt.nz)
At the same time, the national ED scribe announcement said the next tranche of more than 1,000 licences would be used predominantly by mental health teams, suggesting that AI-assisted documentation is likely to become part of routine crisis and frontline mental health workflows before more ambitious AI use cases arrive. (beehive.govt.nz)
Research, Innovation, and the Emerging Pipeline
HRC has created a real national portfolio
The HRC’s annual report confirms that New Zealand invested in 10 studies worth $4.6 million through the 2025 AI in Healthcare RFP. That portfolio includes projects in radiology, stroke, heart failure, dementia, youth mental healthcare ethics, digital pathology, medication safety in older adults, and postoperative monitoring. (hrc.govt.nz)
The portfolio is notably implementation-focused
Several funded projects are not “AI for AI’s sake”; they are explicitly about workflow redesign, feasibility, equity, and real-world deployment:
- Radiology: compare chest x-ray reporting accuracy and turnaround with and without AI, then study barriers to national rollout. (hrc.govt.nz)
- Stroke: test whether AI-assisted scan interpretation increases intervention rates and reduces treatment delay, especially across smaller hospitals. (hrc.govt.nz)
- Heart failure: explore how an AI-derived management support tool could work reliably inside secure digital health systems. (hrc.govt.nz)
- Postoperative monitoring: use digital monitoring and AI to improve early warning of deterioration, with explicit attention to privacy, equity, and Māori data sovereignty. (hrc.govt.nz)
- Youth mental healthcare: co-design guidelines for safe, fair, and culturally responsive AI use with young people, Māori communities, clinicians, and government. (hrc.govt.nz)
Healthy-ageing collaboration with Singapore broadens the agenda
In July 2025, the Government announced a $24 million investment in joint NZ-Singapore research programmes, including an AI stream for healthy ageing. The healthcare-related AI projects include:
- AI-Assisted interRAI Assessment led by the University of Otago
- AI-Driven Risk Score for Dementia led by the University of Auckland
- AI-Augmented Cognitive Health Monitoring led by Victoria University of Wellington. (beehive.govt.nz)
Otago’s follow-up reporting adds that the interRAI project alone will receive $4 million and is intended to partially automate current aged-care assessments so health professionals can spend more time with patients and reduce waiting lists. (otago.ac.nz)
Governance, Regulation, and Trust
Governance is becoming more formal and visible
Health NZ’s National Artificial Intelligence and Algorithm Expert Advisory Group now has a public-facing process and published terms of reference. The group oversees AI initiatives using Health NZ data, evaluates them across ethical, technical, clinical, legal, equity, and operational dimensions, and requires that all AI development or implementation plans be registered with the group. Its remit explicitly includes Māori health and Māori data sovereignty. (info.health.nz)
This matters because NZ is not treating healthcare AI as a generic IT procurement issue. It is building a health-specific oversight mechanism that sits between innovation and operational approval. (info.health.nz)
Privacy controls are shaping the adoption model
Health NZ’s privacy statement says that when it uses AI, it operates in a closed environment so personal information does not leave Health NZ technology and cannot be used to build commercial AI or generative AI systems. Where there is short-term disclosure to AI providers, the statement says personal information is deleted shortly after disclosure and is not used to train third-party commercial models. (info.health.nz)
This strongly suggests the NZ public-sector model will favour:
- closed or tightly governed enterprise deployments
- local validation and approval
- restricted use of general-purpose consumer AI in patient workflows. (info.health.nz)
Professional guidance is catching up
The Medical Council of New Zealand opened consultation in September 2025 on a draft statement, Using artificial intelligence (AI) in patient care, describing what Council expects when doctors use AI in direct patient care. As of early March 2026, search results still show this as a consultation/draft-stage item rather than a final published standard. (mcnz.org.nz)
Trust, equity, and social licence remain central
The Ministry of Health’s precision health programme says NZ is still in the early stages of assessing safe AI adoption and is working to manage risks including discrimination, privacy issues, and the possibility that costs widen existing health gaps. Separately, the AI in Health Research Network explicitly positions ethics, equity, and data sovereignty at the centre of responsible AI translation in Aotearoa. A recent New Zealand Medical Journal viewpoint likewise argues that patient trust will depend on good governance over data and visible Māori representation in AI development and implementation. (health.govt.nz)
Key Patterns and Trends
1. Admin-first adoption is now an established reality
The biggest practical gains are coming from ambient documentation, note generation, summaries, and referral drafting, not from autonomous diagnosis. This is true in public hospitals and primary care alike. (beehive.govt.nz)
2. Imaging is the next serious scale opportunity
Breast screening, stroke imaging, chest x-ray reporting, and diabetic retinal screening all point in the same direction: imaging is where NZ sees the strongest near-term opportunity to relieve workforce pressure while preserving clinician oversight. (beehive.govt.nz)
3. New Zealand is building an adoption model around oversight, not deregulation
The national AI strategy takes a generally enabling approach across the economy, but healthcare adoption is still being filtered through health-specific review, privacy controls, and implementation research. In practice, NZ healthcare is not pursuing “move fast” AI; it is pursuing guardrailed augmentation. (beehive.govt.nz)
4. Equity and data sovereignty are not peripheral issues
Across Health NZ governance, the Ministry’s precision health framing, the AI in Health Research Network, and multiple HRC-funded studies, Māori health, bias mitigation, and data sovereignty appear as recurring design requirements rather than optional add-ons. (info.health.nz)
5. The translation gap is narrowing
A year ago, much of the conversation was still about pilots and principles. By March 2026, NZ has a clearer continuum:
- operational deployment in ED scribes
- active pilots in screening and workflow support
- structured national research funding for implementation
- formal oversight for approval and governance. (beehive.govt.nz)
Overall Assessment
New Zealand healthcare is no longer merely “AI-curious.” It has entered a phase of targeted operationalisation. The strongest evidence is the nationwide ED rollout, the maturing primary care scribe market, and the growing imaging and screening pipeline. Yet the system is still correctly characterised as early-stage overall, because most clinical AI beyond documentation remains in validation, pilot, or funded-research form rather than routine national service delivery. (beehive.govt.nz)
The strategic shape of adoption is now visible: New Zealand is using AI first where it can relieve workforce pressure, improve throughput, and standardise repetitive tasks, while maintaining clinician accountability and a strong emphasis on privacy, equity, and Māori data sovereignty. If that pattern continues, the next 12 to 24 months are likely to bring broader deployment in mental health documentation, screening and radiology support, and workflow-aware primary care copilots rather than abrupt moves toward autonomous clinical AI. (beehive.govt.nz)