Māori are named in only about a third of New Zealand news stories about Artificial Intelligence (AI). Māori are talked about, but rarely heard. When you count who is quoted, Māori technology, rights and data sovereignty voices make up just 3% of media articles.
This creates an issue where Māori societal harms, opportunities and benefits are ignored, creating a media cultural assimilation of Māori and Artificial Intelligence.
Mapping 18 months of AI coverage in New Zealand news media, with no AI-specific legislation and a light-touch national AI strategy, the media is one of the main places our public debate about AI governance and safety actually happens.
This is a clear example of cultural assimilation by media and a handful of academic and industry spokespeople who speak about AI to the media.
What was analysed
The dataset covers 183 news and tech-media articles on AI from 27 New Zealand outlets, published between April 2025 and October 2026. It includes general AI coverage and adjacent areas: facial recognition, automated decision-making, digital identity, biometrics, AI safety, cyber safety and digital harm.
For each article it was recorded who wrote it, who was quoted, and whether it mentions Māori. That gave 334 contributor appearances by 216 people and organisations. Every Māori code was then checked by a human (the author), because automated reading of search excerpts badly undercounts Māori mentions.
Findings
1. Māori are named, but rarely heard
Māori are mentioned in 59 of the 183 articles (32%). But of 334 contributor appearances, only 11 were Māori technology, rights or data sovereignty voices. Academia and Industry voices outnumber them eight to one. Five of those 11 appearances are mine. I say that not as a boast but as a warning. When one person carries that much of the load, a Māori position on AI starts to look like one expert’s opinion. We already see this with the Māori Data that many of the public AI has been trained on.
2. Māori appear as a population at risk
Where Māori are named depends heavily on the technology. Facial recognition and automated decision-making stories named Māori in 11 of 14 articles (79%), compared with 31% of general AI stories.
3. Silence on cyber safety, where Māori face more harm
Not one of the 13 articles on AI safety and cyber safety mentioned Māori. Digital harm stories, covering deepfakes and image-based abuse, did so in only 3 of 13, despite Māori being disproportionality victimised with online harm.
The InternetNZ 2025 Internet Insights survey, which shows Māori are more exposed to online harm and more worried about it than NZ Europeans:
- Personally experienced online harm or harassment: Māori 20%; NZ European 14%.
- Concerned about personal data security: Māori 61%; NZ European 43%.
- Concerned about identity theft: Māori 70%; NZ European 57%.
- Concerned about online crime: Māori 69%; NZ European 61%.
- Concerned about cyberbullying: Māori 69%; NZ European 63%.
- Turn to friends and whānau for online safety advice: Māori 34%; NZ European 23%.
Māori also report less knowledge of AI (23% say they know a fair amount or a lot, against 31% overall) and more concern about it (54% against 47%).
None of this is new. I have been tracking Māori online safety data since 2018, and the pattern has held for most of a decade. Below are some statistics with many more Te Kete o Karaitiana Taiuru Blog:
- 2018: NetSafe’s research on teens found 16% of Māori teens had been threatened online, against 11% of Pākehā and 8% of Pacific teens. Māori girls aged 14 to 15 were the most vulnerable group.
- 2018: In NetSafe’s online hate speech study, 90% of Māori who had experienced online hate believed it was because of their ethnicity.
- 2019: NetSafe’s image-based abuse research found threats to share intimate images were more common among Asian 8% and Māori 6% respondents.
- 2021: NetSafe’s Global Kids Online parenting study under-sampled Māori and Pacific whānau and relied on post-survey weighting to make up the gap.
- 2022: The Classifications Office found Māori were more often exposed to online content promoting racial violence, violent extremism and racial hatred. Only 24% of Māori strongly agreed they knew enough to help whānau stay safe online.
- 2026: Office of the Privacy Commission found New Zealanders, 52% say they are concerned about their individual privacy, Māori 57%.
- 2026: Office of the Privacy Commission found 58% of Māori are concerned about biases in facial recognition technology, compared to 53 percent overall.
- 2026: Office of the Privacy Commission 21% of Māori have very high or high trust in government agencies to handle their personal information responsibly compared to 31% of NZ.
The evidence has been public for years. The research has often sampled Māori poorly, and the news coverage has rarely picked it up. Yet stories on cyber safety still treat risk as falling on an undifferentiated public. They never ask whether the harm, or the remedy, is different for Māori.
4. Māori media carry the load
Waatea News named Māori in every one of its AI articles and supplied almost a fifth of all Māori mentions in the dataset. Mainstream and tech AI reporting does not build Māori perspectives in.
The specialist newsletter The AI Corner shows the same dependency. Its weekly round-ups of other outlets’ reporting named Māori in 42% of posts, while its own commentary did so in just 4%. Its Māori mentions also fell sharply, from 58% of round-ups in 2025 to 24% in 2026.
The harms of leaving Māori out
I often ask whether Māori provisions in a policy are architecture or decoration: do they actually shape how the thing works, or are they there for appearance? The same test applies to news. A passing mention of Māori is decoration. Architecture is when Māori are positioned as authorities who define the problem and its remedy. On the evidence of this dataset, It tells the public everything is fine.
Media shape what people think about, not just what they think. If Māori concerns about AI rarely surface, the public and decision-makers can reasonably conclude there are none.
The safety silence worries me most. AI safety, deepfakes and online harm are where new regulation is now being written, including the deepfakes bill. Māori already experience online harm at higher rates. If we are absent from how the public understands these risks, we will be absent from how the problem is defined in law, and the remedies will be designed for someone else. I have argued before that New Zealand’s voluntary responsible-AI instruments leave Māori governance outside their structure. Our media coverage is reproducing that exclusion in public debate.
It starves Māori-led work of funding. Māori-led AI initiatives, research and safeguards compete for funding against issues that are far more visible. Māori are developing AI tools, governance frameworks and data protections. When these go unreported, they are not evaluated, resourced or scaled.
Te Tiriti o Waitangi should be a constitutional threshold that AI decisions must clear, not something added once the technology is chosen. Right now our public conversation treats it as the afterthought.
What needs to change
For media:
- Seek out Māori technology, rights and data sovereignty experts as sources on every AI story, not only stories labelled Māori.
- Go beyond the one or two Māori experts you already know, and seek out community voices directly.
- Ask the Te Tiriti question in every AI policy story, and bring Māori harm, safety and data sovereignty to AI stories as a frame rather than waiting for an official document to raise them.
- One in five Māori has experienced online harm, and a third turn first to friends and whānau for help. Stories on scams and deepfakes should reflect that, and should recognise the role Māori institutions can play in the response.
For researchers: the next steps are to code whether Māori are mentioned in passing or treated substantively, to track Te Tiriti references across all outlets, and to separate community voices from individual experts.
For regulators, funders and policymakers: silence in the coverage is a gap in the evidence, not proof that Māori have no concerns. Media coverage is no substitute for engagement with Māori as Treaty partners. If the public debate is leaving Māori out, that is a reason to make Te Tiriti obligations structural in AI governance, not a reason to assume the issue has been aired.
The absence of Māori voices is not a neutral omission. It shapes what the public believes, what regulators prioritise and where resources go. The question is not whether Māori have something to say about AI. It is why the media isn’t asking.
Read the full paper
The full methods and findings paper, Mapping AI Coverage in Aotearoa New Zealand News Media, sets out how the dataset was built, every count behind these findings, and the limits of what they can show. [LINK TO PAPER]
The dataset is designed to be checked, challenged and extended. If you are a journalist, researcher or editor who wants to help close the gaps, I would like to hear from you. Māori deserve to be heard in the debate about the technologies that will shape our future, not just named in it.