Facial Recognition Technology, Biometrics and Digital Identity – Collected Posts
Dr Karaitiana Taiuru · compiled 11 October 2026
Twelve posts by Dr Karaitiana Taiuru, published December 2020 to March 2026, track facial recognition technology (FRT) in Aotearoa and its risks to Māori: bias, tikanga breaches, and Māori data sovereignty. For a background of these technologies see the FAQ’s page.
Read our glossary for more AI and Māori Sovereignty terms.
Dates are original publication dates as shown on each post. Every post also shows a “last updated” date of 7–9 October 2026, which appears to be a site-wide update, so those dates are not used here. Summaries are paraphrased; follow each link for the full post. A further 37 posts from the same category, on digital identity, DNA and wider AI governance (May 2017 to September 2026), are listed under Related posts at the end, with one-line summaries drawn from the category listing. All posts except Māori stats re AI, FRT and Privacy (filed under AI Governance and Statistics) sit in the Digital Identity Facial Recognition & Biometrics category on taiuru.co.nz.
Timeline at a glance
Posts with summaries, oldest to newest
Māori Cultural considerations with Facial Recognition Technology in New Zealand
The foundational piece, sixth in a series on Māori ethics with AI. It likens FRT to the colonial collection of mokomokai and maps the FRT already held by Police, Customs, Justice and DIA, with little regulation or engagement with Māori. It sets out ten Māori-specific issues, including moko as taonga, images stored offshore, and the cultural harm of storing images alongside those of the dead. Eight recommendations build on Victoria University’s 2020 FRT report: a Te Tiriti clause in every government FRT system, a biometrics code co-written with Māori and iwi, a Biometrics Commissioner with Māori seats, cultural audits, and an opt-out scheme with removal of images of the deceased. It also names positive uses, such as identifying repatriated mokomokai or unknown tīpuna in old photographs. Read the post
Māori being used as ‘guinea pigs’ for facial recognition tech by MSD
Expands on an RNZ interview about MSD’s rollout of FRT for beneficiaries to verify identity online. The post argues MSD proceeded without a privacy, human rights or ethics framework, and without consulting Māori. RNZ had reported the government system failed about half the time in tests, and OIA documents suggested it could fail more often for Māori. It flags Google Analytics sending metadata to the US as inconsistent with the WAI 2522 finding that Māori data is a taonga. It frames the rollout as a Te Tiriti and human rights breach, comparable in intergenerational impact to the Hunn Report. Read the post
Facial Recognition and Artificial Intelligence Profiling
Commissioned by Ngā Toki Whakarururanga for a general audience, and updated for Foodstuffs North Island’s supermarket trial in seven North Island stores. It reviews international evidence (Gender Shades, NIST) that FRT is least accurate for darker-skinned women, and New Zealand evidence including the Police’s halted FRT use and the failed Clearview AI trial. It warns that Māori, Pasifika and women shoppers will bear false matches, that staff tend to over-trust algorithmic matches, and that trespassing “associates” ignores whānau shopping together. It also covers moko identity theft, offshore storage under foreign law, and FRT-driven advertising. Read the post
First Māori woman victim in facial recognition trial in a month
Five weeks after warning that Māori women would be misidentified, Dr Taiuru reports that a Māori mother in Rotorua was wrongly flagged as a trespassed thief at a New World store. Staff rejected her photo ID despite Foodstuffs’ assurances of trained verification at a 90% match threshold. The post blames automation bias, with staff trusting the match over the evidence, and asks whether racial stereotyping played a part. It proposes fixes: staff with lived ethnic experience, bias training on New Zealand datasets, clearer signage, de-escalation training, senior-manager dispute rules, and links with community leaders. Read the post
Inconsistencies with Facial Recognition Trial
Summarises Consumer NZ’s analysis of Foodstuffs North Island’s trial numbers. Key points: Foodstuffs declined Consumer NZ’s data requests, the misidentification risk was about 1 in 175 (ten times higher than Foodstuffs’ reading), no ethnicity data was collected, and the claimed reduction in assaults may be overstated. The Privacy Commissioner was then evaluating the trial. The lack of transparency heightens concern for Asian, Pacific, Indian and Māori shoppers. Read the post
Facial Recognition trial report obfuscates risks to Māori
Critiques the BixeLab bias evaluation of DIA’s Identity Check system. The test used 148 DIA staff against a recommended 245 for 95% confidence, did not separate Māori from Pacific peoples, and included no one with facial moko, although the report acknowledged them as 10–30% of the demographic. Impostor-match data still pointed to higher risk for Māori and Pasifika, and for moko wearers. Recommendations: test the full sample, record Māori participants by appearance and moko, and add a Māori cultural expert as peer reviewer. RNZ’s follow-up prompted DIA to commit to further studies. Read the post
NZ Government Facial Recognition Technology
Responds to Newsroom’s report that the Ministry of Justice reviewed the Privacy Act in 2024 for barriers to retail FRT, without Māori, ethnic or privacy experts, and withheld the results. It notes the retail crime Ministerial Advisory Group has no Māori members. Proposed solutions: mandatory, renewable FRT ethics and safety certification for operators, and a sovereign New Zealand FRT system built from driver licence and passport photos and licensed to retailers, meeting Te Tiriti requirements. Read the post
Introduction to Facial Recognition bias
A free resource for whānau Māori, with recommended documentaries (including Coded Bias), a podcast, DIA’s Biometrics 101 course, and a media reading list. It recaps the Rotorua case, Consumer NZ’s critique, and the DIA moko findings. It criticises the Privacy Commissioner’s consultation timing over the 2024–25 summer break and the make-up of its Māori advisory group, which it says lacks AI, privacy and FRT expertise. Read the post
Māori stats re AI, FRT and Privacy
Pulls Māori results from the Privacy Commissioner’s March 2025 survey. Māori concern exceeded the general population on FRT bias (57% vs 48%), police FRT (52% vs 41%) and retail FRT (49% vs 41%), even as overall concern fell. 86% of Māori want more say over their data, and 19% reported a privacy breach, up 7% from 2022. The post concludes FRT mistrust will persist until systems are trained on a New Zealand dataset. Read the post
Customary Māori Values and Copyright Legislation for Facial Recognition Technology (FRT)
Written after the Privacy Commissioner found Foodstuffs’ trial complied with the Privacy Act. It notes Māori community consultations were not opposed to FRT outright but wanted co-design, tikanga, human oversight and secure storage. It sets out tikanga about the face, reflections and shadows. It proposes a New Zealand law, modelled on Denmark’s planned deepfake law, letting people hold copyright in their face, voice and biometrics, with protections for the deceased and moko. Read the post
AI Facial Detection Technology miss-identified 3 Māori men
Documents Google’s AI search mode presenting photos of Brian Dickey KC and Kingi Snelgar as Dr Taiuru. Google says this is face detection and visual similarity, not facial recognition, but the effect is similar. The post calls for human checking of AI outputs and specialised bias training for anyone verifying FRT matches. Read the post
Facial Recognition, Algorithmic Bias, and Compounded Harm to Māori and Pacific Communities
Introduces a paper on Foodstuffs South Island’s Christchurch trial. The three trial stores are in suburbs with well-below-average Māori and Pacific populations, while the suburbs with the highest are excluded. International research shows false positives up to 100 times higher for Indigenous and Pacific peoples. The Privacy Impact Assessment has no New Zealand bias testing, no ethnicity monitoring, and no Māori data sovereignty consultation. Calls for independent bias testing, transparent reporting and real Māori partnership in governance before expansion. Read the post · Full paper
Related posts in the category
The remaining 37 posts in the category, newest first within each theme. These summaries come from the category listing excerpts, not full readings of each post.
Digital identity, images and deepfakes
| Published | Post | Focus |
|---|---|---|
| 28 Sep 2026 | Principles for Digital Identity | Māori identity in digital ID systems as relational and collective, grounded in whakapapa, not one person per account |
| 21 Jun 2026 | Deep fake harms real scenario and new legislation | A local deepfake harm case and new legislation |
| 17 Jun 2026 | Tikanga AI images of a person | Huffer allegedly generating AI models from past campaign models’ likeness, viewed through tikanga |
| 19 Mar 2026 | Meta’s Digital Ghost Patent | Te Ao Māori analysis of Meta’s patent to simulate users from their data |
| 3 Dec 2025 | Māori views of Digital Credentials | Māori perspectives on digital credentials and wallet apps |
| 12 Oct 2025 | AI and Pepeha | Māori disconnected by assimilation turning to AI to reclaim identity and pepeha |
| 13 Apr 2025 | Risks of AI Action Figure Trend | Privacy, appropriation and deepfake risks of AI avatar trends |
| 20 Sep 2021 | Tikanga with face swapping filters and deep fakes | Tikanga on face swaps and deepfakes, also relevant to FRT |
| 23 Mar 2021 | Origins of the term Pākehā | Historical analysis of the word Pākehā |
| 20 Mar 2021 | Tikanga about animated pictures of the dead | Tikanga breaches when apps animate images of tūpuna |
| 3 Oct 2020 | Māori business self identification | Option for Māori businesses to self-identify on the NZBN register |
| 15 Jul 2017 | Iwi Citizen or Tribal Member | Why “tribal member” is a colonial term |
| 28 May 2017 | Protecting digital ancestor images | Tikanga on photographing and sharing images of tūpuna |
DNA, genetics and biological data
| Published | Post | Focus |
|---|---|---|
| 8 Dec 2025 | Māori Forensic Deceased Data | Dr Angela Clark on digital governance of tūpāpaku data |
| 1 Nov 2024 | Te Ao Māori perspective of Privacy with DNA | Cultural and privacy risks of commercial DNA testing |
| 20 May 2022 | Tikanga Tawhito Tikanga Hou | Full bioethics guideline for Māori genetic samples (ISBN 978-0-9582597-8-1) |
| 20 May 2022 | Tikanga Associated with Biological Samples | Tikanga for research and storage of Māori biological samples |
| 20 May 2022 | Indigenising DNA | DNA explained through mātauranga Māori and whakapapa |
| 13 May 2022 | Tikanga Tawhito Tikanga Hou – Literature Reviews | Review of existing Māori ethics frameworks for DNA |
| 2 Jan 2021 | Maori Cultural DNA | Law Commission’s 2020 DNA report and its Te Tiriti recommendations |
| 22 Mar 2019 | Submission to the Law Commission: DNA in Criminal Investigations | Public submission on DNA as taonga |
| 11 Feb 2019 | DNA is a Taonga: A Customary Māori Perspective | Paper on DNA as taonga, written for the Law Commission consultation |
| 29 Jan 2019 | Law Commission DNA consultation: Māori implications | Notice of the consultation on the CIBS Act 1995 |
| 2 Aug 2018 | Introduction to Māori Cultural Considerations with Genetics | Tikanga and cosmology applied to genomic research |
| 6 Jun 2017 | Dangerous game of DNA testing for Maori | Scenario analysis of DNA ancestry testing risks |
Digital Identity
| Published | Post | Focus |
|---|---|---|
| 3 May 2026 | AI Agents and Legal Personhood in New Zealand | Mead’s tikanga framework applied to agentic AI and personhood |
| 15 Apr 2026 | Google DeepMind hires ‘philosopher’ | DeepMind’s hire of a philosopher to study machine consciousness |
| 14 Apr 2026 | Personal Māori Data Sovereignty | Practical privacy steps for whānau on Windows computers |
| 26 Mar 2026 | AI Bias with Māori in Job Interviews | AI recruitment bias and media neglect of Māori voices |
| 6 Feb 2026 | Autonomous AI Governance Guidance for New Zealand Organisations | Governance guidance for deploying autonomous AI agents |
| 26 Oct 2025 | Risks of AI Agents to Māori | AI agents as a new wave of colonisation risk |
| 5 Nov 2022 | The NZ Supreme Court Judgement and the impacts to Māori Data Sovereignty | Tikanga as common law and what it means for data sovereignty |
| 20 Mar 2022 | Consumer Guidelines for Māori Non Fungible Tokens | Appropriation risks in NFTs |
| 22 Feb 2022 | Māori Data Sovereignty and Domain Names | Who can use the maori.nz domain |
| 30 Oct 2021 | Marae Data Sovereignty Data Sets | Open datasets of marae names for marae data sovereignty |
| 8 Nov 2020 | Māori Data Sovereignty and Associated Legal Instruments | Definitions of Māori data and Crown obligations, since folded into the Compendium |
| 19 Dec 2018 | Data is a Taonga. A customary Māori perspective | Foundational paper on data as taonga |
How to cite this page
Taiuru, K. (2026, September 20). Biometrics. Taiuru & Associates Ltd. https://www.taiuru.co.nz/biometrics/