Latest

AI in Academic Research in Aotearoa New Zealand: A Living Whitepaper

Last updated: 9 March 2026

Introduction

AI adoption in New Zealand’s academic research sector has moved beyond early experimentation. Since mid-2025, the strongest signals have been institutional: a national AI strategy, a government-backed AI research platform process, new digital research infrastructure, research-funder guidance, ethics-review requirements, and university-level governance documents. The result is a sector that is not yet uniformly mature, but is clearly shifting from informal tool use toward structured, governed, and mission-led adoption. (mbie.govt.nz)

Executive Summary

  • The sector’s center of gravity is moving from curiosity to capability. The biggest step-change came on 8 July 2025, when MBIE released New Zealand’s first AI strategy; this was followed by the NZIAT call for an AI Research Platform with funding of up to NZ$70 million over seven years to strengthen research capability and commercialisation. (mbie.govt.nz)
  • Health research is the clear frontrunner. In 2025 alone, the Health Research Council launched a dedicated AI in Healthcare funding initiative with a NZ$5 million pool, and later reported 10 funded studies worth NZ$4.6 million. In parallel, MBIE funded three New Zealand–Singapore “Leveraging AI for Healthy Ageing” projects, each with NZ$4 million for the New Zealand partner. (gateway.hrc.govt.nz)
  • Governance has become much more explicit. Royal Society Te Apārangi’s June 2025 guidance, Marsden Fund rules, HDEC supplementary AI forms, and new university policies all emphasize human accountability, disclosure, confidentiality, equity, and Māori data sovereignty. (royalsociety.org.nz)
  • Infrastructure is improving. On 5 August 2025, REANNZ and NeSI were integrated into a unified eResearch platform to support high-performance computing, AI capability development, and access to research data networks. Universities are also standardising secure AI access for staff and researchers. (mbie.govt.nz)
  • The sector still lacks a clean baseline measure of adoption. One of the most notable signs of this gap is the national tertiary-sector survey launched by Massey University on 19 March 2025 to gather baseline evidence on AI use in teaching, research, and professional work. (ako.ac.nz)

Current State of AI Adoption

High-level snapshot

AI adoption in New Zealand academic research is currently best described as selective, fast-rising, and unevenly distributed. Adoption is strongest where there are large datasets, high-value prediction problems, clear translational pathways, and access to specialist compute or clinical partners—especially in health, environmental science, engineering, and computer science. (gateway.hrc.govt.nz)

What is changing fastest is not just the number of research projects, but the institutional scaffolding around them:

What has changed since the placeholder stage

Since the earlier placeholder article, several concrete developments have materially changed the picture:

  • 8 July 2025: MBIE released the country’s first AI strategy. (mbie.govt.nz)
  • 5 August 2025: MBIE announced a unified national eResearch platform through REANNZ–NeSI integration. (mbie.govt.nz)
  • Late 2025 to early 2026: universities including AUT and Otago issued more formal AI research governance documents. (aut.ac.nz)
  • 18 December 2025: a University of Auckland-led proposal for a national Agentic AI Platform received seed funding for detailed development. (auckland.ac.nz)

Taken together, these developments suggest that New Zealand academic research is moving from isolated AI projects toward a more coordinated national research system. That is an inference, but it is strongly supported by the convergence of policy, funding, infrastructure, and institutional governance changes across 2025–2026. (mbie.govt.nz)

Latest News and Strategic Developments

1) National policy now explicitly backs AI-enabled research

MBIE’s AI strategy, released on 8 July 2025, positions AI adoption as a national economic priority and links that agenda directly to the science, innovation, and technology system. The strategy says AI will be supported through mechanisms including the Marsden Fund, the Catalyst Fund, and the creation of a new advanced-technology public research organisation with AI in scope. It also notes that AI-related projects approved through the R&D Tax Incentive had an estimated spend of NZ$611 million since 2019. (mbie.govt.nz)

2) A national AI Research Platform is being built

MBIE’s Artificial Intelligence Research Platform call states that the government is investing to strengthen New Zealand’s research capability and commercialisation in AI, with up to NZ$70 million over seven years available in the platform-selection phase. The platform is expected to deliver internationally leading AI research that creates new firms and opens new market possibilities for New Zealand. (mbie.govt.nz)

On 18 December 2025, University of Auckland researchers announced that their proposal, “Aotearoa Agentic AI Platform: A Productivity Multiplier,” had received NZ$250,000 in seed funding. The proposed platform’s initial focus areas are healthcare, robotics and automation, and science discovery including drug discovery. (auckland.ac.nz)

3) Digital research infrastructure is now being aligned for AI

MBIE announced on 5 August 2025 that REANNZ and NeSI had integrated to form a unified eResearch infrastructure platform. The platform is intended to support high-performance computing, AI capability development, specialist data-network access, and researcher training, backed by NZ$13.93 million per year through the Strategic Science Investment Fund. (mbie.govt.nz)

Research, Governance, and Institutional Readiness

National research guidance is now more mature

Royal Society Te Apārangi’s June 2025 guidelines for the best-practice use of generative AI in research are now the clearest national reference point. The guidelines state that researchers remain accountable for AI-assisted content, must maintain a critical stance toward outputs, must disclose substantial AI use, and must not list GenAI systems as authors or co-authors. They also explicitly flag cultural limitations, bias, hallucinations, and environmental impacts. (royalsociety.org.nz)

The Marsden Fund has also adapted. Its 2025 applicant guidelines warn that generative AI can raise authorship, intellectual-property, and factual-accuracy issues, and require applicants to take full responsibility for proposal content. Separate guidance for Marsden panellists and reviewers says they must not use LLMs or other generative AI tools in proposal assessment because of confidentiality risks. (royalsociety.org.nz)

Ethics oversight is tightening

Since 30 August 2024, Health and Disability Ethics Committees have required a supplementary form when AI or algorithmic tools are used or developed as part of a research project, with reference to Chapter 13 of the national ethical standards on health data and new technologies. Applications with AI components can be sent back if the form is missing. (ethics.health.govt.nz)

The HRC’s 2025 AI in Healthcare application guidelines reinforce the same direction: proposals are expected to address ethical, safety, data-governance, and sovereignty matters, and show how the work will improve health outcomes in New Zealand. (gateway.hrc.govt.nz)

Universities are formalising AI use in research

Recent university signals show a move from permissive experimentation to governed adoption:

  • AUT’s January 2026 research guidelines emphasise participant-data protection, disclosure when AI materially affects the research process, and strong treatment of Māori and Pacific data sovereignty, including alignment with Te Mana Raraunga and the CARE Principles. (aut.ac.nz)
  • University of Otago’s AI Governance Policy, approved in December 2025 and taking effect on 10 March 2026, states that AI governance across university activities must align with Te Tiriti obligations, data sovereignty, integrity, transparency, and accountability. (otago.ac.nz)
  • Massey University’s staff guidelines say GenAI can support research by fostering innovation and accelerating discovery, while also stressing privacy, security, equity, Māori data sovereignty, and research conduct risks. (massey.ac.nz)
  • University of Auckland now provides staff and students with approved access to Microsoft 365 Copilot Chat, Google Gemini, and NotebookLM, explicitly positioning NotebookLM as an AI-powered research assistant for working with users’ own documents and sources. (auckland.ac.nz)

Scholarly publishing norms are shifting too

The New Zealand Journal of Employment Relations now requires authors to disclose substantive AI use in the research process and states that AI cannot be listed as a co-author. It also bars reviewers from uploading manuscripts into AI tools because of confidentiality and IP concerns. This is a small but important sign that AI governance is now extending into New Zealand’s scholarly communication layer. (ojs.aut.ac.nz)

Research and Case Studies

Case Study 1: AI for healthy ageing has become a national research cluster

MBIE’s Catalyst: Strategic – New Zealand-Singapore Leveraging AI for Healthy Ageing 2025 created one of the most concentrated AI research pushes in the country. The programme funds three projects, each for three years starting in mid-2025, with NZ$4 million to each New Zealand partner. (mbie.govt.nz)

The three projects illustrate how academic research is using AI in translational ways:

  • University of Auckland: an AI-driven dementia risk score that combines brain, blood, cognitive, genetic, clinical, and general health data, with explainable AI and co-design for clinical use. (mbie.govt.nz)
  • University of Otago: AI-InterRAI, which aims to partly automate interRAI assessments, cut assessment time by 50 percent, improve risk prediction, and generate AI-driven personalised care plans. (mbie.govt.nz)
  • Victoria University of Wellington: AIMCura, a mobile-first, AI-augmented platform using speech tasks and cognitive games for remote monitoring, risk assessment, and personalised cognitive support. (mbie.govt.nz)

Case Study 2: HRC funding shows breadth beyond a few flagship labs

The HRC’s 2025 AI in Healthcare initiative created a broader pipeline of academic and translational work, with a total funding pool of NZ$5 million and later-reported investment of NZ$4.6 million across 10 studies. (gateway.hrc.govt.nz)

Notable examples include:

  • AI-enabled diabetic retinopathy screening, led from Health NZ Waitematā, testing AI for faster and more equitable eye-disease screening, especially for Māori and Pacific communities. (hrc.govt.nz)
  • AI-assisted acute stroke care, aiming to improve rapid radiological interpretation; if successful, it is expected to enable an additional 850 New Zealanders per year to access critical treatment. (hrc.govt.nz)
  • AI-enhanced digital pathology at the University of Otago to improve gastrointestinal cancer prognosis and treatment decisions. (hrc.govt.nz)
  • Explainable AI for dementia risk factors at the University of Auckland, using the Stats NZ Integrated Data Infrastructure to identify modifiable risks while focusing on transparency and inequity reduction. (hrc.govt.nz)
  • Guidelines for AI in youth mental healthcare at AUT, explicitly co-designed with young people, Māori communities, practitioners, and government agencies. (hrc.govt.nz)

This portfolio matters because it shows AI adoption in academic research is not confined to model-building alone; it increasingly includes implementation, evaluation, workflow integration, and equity design. (hrc.govt.nz)

Case Study 3: Environmental and engineering research is adopting AI for field-scale problems

At the University of Canterbury, a 2025 programme in aquaculture is using AI and 3D imaging with autonomous underwater vehicles to inspect shellfish and seaweed systems. UC says the work could deliver immediate benefits of up to NZ$80 million per year to the mussel industry alone. (canterbury.ac.nz)

At Earth Sciences New Zealand / NIWA, researchers developed a generative AI approach for high-resolution climate projections that the organisation says is more than 1,000 times faster than current physics-based methods. In 2025 it released the first AI-produced REMS-MR downscaled dataset, covering more than 15,000 years of model simulations across multiple climate models and emissions scenarios. (niwa.co.nz)

These examples show a second major adoption pattern in New Zealand academia: using AI where physical systems, sensors, and simulation create scale problems that traditional workflows struggle to handle. (canterbury.ac.nz)

Case Study 4: Science discovery and digital twins are moving onto the agenda

A third frontier is AI for science discovery itself. The University of Auckland-led Aotearoa Agentic AI Platform proposal explicitly includes science discovery and drug discovery among its first focus areas. (auckland.ac.nz)

Separately, on 18 September 2025, MBIE announced up to NZ$4.5 million over four years for a partnership between the Auckland Bioengineering Institute and the University of Texas at Austin to develop AI-enabled digital twins for health and bioeconomy applications. MBIE said these models could support personalised healthcare, accelerate drug development, and generate both hospital savings and licensing revenue. (mbie.govt.nz)

1) Health is the leading edge, but not the whole story

The densest concentration of funding, governance work, and cross-institution collaboration is in health. That includes HRC grants, Catalyst funding, ethical evaluation frameworks, and the AI in Health Research Network supported by the University of Auckland and MedTech-iQ. (gateway.hrc.govt.nz)

But the broader pattern is increasingly visible in climate science, aquaculture, engineering, digital twins, and science discovery. (canterbury.ac.nz)

2) Adoption is being shaped by New Zealand-specific values

Across national and institutional guidance, AI adoption in research is being framed through Te Tiriti obligations, Māori data sovereignty, equity, and Pacific values, not just through generic international AI risk language. AUT’s 2026 guidance is especially explicit on Māori and Pacific data sovereignty, while Otago’s governance policy embeds Te Ao Māori and Pacific values, and Royal Society guidance highlights cultural limitations and responsible use. (aut.ac.nz)

3) The sector is shifting from tool access to workflow redesign

Secure tool access matters, but the more important development is workflow redesign: AI-assisted assessment and care planning, AI-enabled screening, climate-model emulation, digital pathology, AI-powered mobile monitoring, and science-discovery platforms. This suggests New Zealand academia is beginning to use AI as a research infrastructure layer, not just as a writing or summarisation aid. That is an inference from the current project mix, but it is well supported by the funded use cases now on record. (auckland.ac.nz)

4) Measurement still lags activity

There is strong evidence of activity, but limited sector-wide quantification of researcher adoption across all disciplines. The Massey-led tertiary survey is important precisely because it aims to produce baseline evidence on AI use in teaching, research, and professional work across institutions. (ako.ac.nz)

Overall Assessment

As of 9 March 2026, AI adoption in New Zealand academic research is best understood as institutionalising rather than saturating. The sector is not uniformly AI-native, and robust nationwide measurement is still incomplete. But the strategic direction is now unmistakable: government has put AI into national research and innovation settings; funders and ethics bodies have created AI-specific processes; universities are formalising secure use and governance; and flagship research activity is expanding from health into environmental science, engineering, and science discovery. (mbie.govt.nz)

The most important takeaway is that New Zealand’s academic research sector is not pursuing AI as a generic productivity fad. Its strongest current pattern is more distinctive: mission-led adoption under explicit governance, with an unusually visible emphasis on equity, clinical translation, data sovereignty, and local relevance. If the AI Research Platform and upgraded eResearch infrastructure deliver as intended, the next phase is likely to be larger-scale, more coordinated, and more internationally visible than the sector’s current project-by-project footprint. (mbie.govt.nz)

Sector updates

Article history

No previous updates have been archived yet.

Back to all sectors