The minutes for each meeting of the Data Ethics Advisory Group from 2026, grouped by date, are summarised here.
Meeting summaries from 2019 to 2022
Members: Professor Colin Simpson (Chair), Andrew Sporle, Jonathan Godfrey, Kate O’Connor, Russell Craig, Will Koning
Apologies: Jonathan Kilgour
Attendees: Centre for Data Ethics and Innovation (CDEI), Stats NZ
Presenters: Stats NZ representatives
Meeting opening and standing items
The Chair noted the apologies. In relation to an agenda item, a member reminded of their financial interest pertaining to their company’s commercial arrangement with Stats NZ. Due to the nature of the interest and its distinction from the agenda topic, it was agreed that there was no conflict of interest and the member remained for the meeting.
Presenters provided context around Stats NZ’s effort to develop an approach for deciding on the use of linked admin data in statistical operations, with a caveat that implications could go wider, including across other agencies. They noted that the approach consisted of two parts: robust organisational policies and Human Values for Data Ethics (HVDE) assessment for individual use.
It was added that the work was in a rough stage and that there were many questions (e.g., strategic, red lines for not using data for any other purpose) and implications for household surveys, more broadly (e.g., improving outcomes and quality for estimates for iwi Māori and other priority communities; improving efficiency of data collecting approach; electoral roll for tactical solutions in the future).
Stats NZ advised that the work was still in an early stage, with unresolved strategic questions and wider implications for household surveys, data quality, efficiency and outcomes for iwi Māori and other priority communities. DEAG members emphasised the need for robust sampling, clear guardrails, social licence, end-to-end data governance, independent oversight and mature escalation pathways. DEAG suggested the need for a foundational, transparent policy rather than ad hoc decisions. DEAG further stressed the need to incorporate Māori data governance, the Mana Ōrite agreement and engagement with Te Kāhui Raraunga. presenters confirmed that a Māori Data Governance overlay would be integrated and that engagement with the Crown Māori Statistical Design Forum was planned.
DEAG outlined the potential conflict between the Government Statistician and Government Chief Data Steward roles and recommended considering an independent oversight model, noting the UK and Australian models were globally leading.
The team gave an overview of the ISDS progress since the team last attended a DEAG meeting, in July 2025:
DEAG members discussed the established governance and data stewardship framework for the use of unit record person data in the ISDS. They were advised that the framework leveraged the IDI best practice, which had an impact on both the ISDS team and its users. DEAG members recommended an ethical governance framework above and beyond the ISDS to address the cultural issue of data handling, acknowledging a range of data protection standards across Stats NZ. They highlighted the importance of a strong centre, a policy for cultural standardisation and for mitigating precedent decision-making. Presenters agreed that ethical considerations should be at the heart of what Stats NZ does, also represented in the operating model and led from the top.
DEAG was provided with a project update, following Cabinet’s approval of $57.7m in funding over four years. It was noted that current users would continue using the IDI, with a plan for upgrades over three years, progressive roll out of improvements in 2030, more frequent data updates and improved data quality, all delivered in a way that would maintain public trust and confidence.
Responding to why IDI needed modernisation, Stats NZ representative advised that Data Labs were aging and restrictive, the current IDI not usable for social investment, and there was a need for growing user base by being more user friendly. It was noted that the Administrative Population Model would be used as a future spine for IDI research (people and household data together), with stewardship being central to the project success.
Presenters proposed methodology for extracting information from names to focus on ethnicity due to issues with Level 4 data. They explained how looking at common patterns in names would contribute to the machine learning model and would not be directly predicting ethnicity.
DEAG members were interested in implications for the Statistical Standard for Ethnicity and were advised on limitations of admin data, hence the initial thinking if the methodology could be used in lost statistical imputation step. It was noted that admin data accuracy across agencies would improve under the Admin Data Improvement Plan and that the Crown Māori Statistical Design Forum would also be consulted.
While noting the approach was academically interesting, DEAG members raised security concerns around its application, advising this methodology should be limited to gaps in existing data, with guardrails in place, especially considering wider system issues with ethnicity data. They were concerned around discrimination and bias, especially with Māori and Pacific people, and Stats NZ setting precedent for other agencies, who would not necessarily adhere to the same standard and could use the methodology inappropriately. For DEAG members, the main considerations were unintended consequences and a potential for real and perceived harm.
The meeting ended with a karakia.
Members: Professor Colin Simpson (Chair), Andrew Sporle, Jonathan Godfrey, Jonathan Kilgour, Russell Craig, Will Koning
Apologies: Kate O’Connor
Attendees: Centre for Data Ethics and Innovation (CDEI), Stats NZ
Presenter: New Zealand Transport Agency representatives
Meeting opening and standing items
The Chair noted the apologies. A member excused themselves from the meeting due to a potential conflict of interest.
DEAG discussed the ethical considerations associated with connected vehicle technology, including Cooperative Intelligent Transport Systems (C-ITS). It was noted that these systems enable communication between vehicles, road infrastructure and other road users, with the potential to improve road safety and traffic flow.
The presenter explained that international studies indicate connected vehicle technology could help mitigate crashes and support safer transport outcomes. They added that, while no related policies had yet been developed in New Zealand, the discussion was intended to seek early ethical guidance as international adoption increases. It was clarified that Connected and Automated Vehicles were not in scope.
DEAG members acknowledged the potential public benefits of the technology, including improved safety for drivers, pedestrians, cyclists and vulnerable road users. Members also noted the importance of considering infrastructure-to-vehicle communication and comparable use cases, such as railway crossings.
Members emphasised that privacy, consent, transparency and public trust would be central to any future use of connected vehicle data. They raised concerns about the potential to identify individuals by combining multiple datasets, such as vehicle, cellphone or licence plate information, and highlighted the need to manage risks of government surveillance or unintended secondary use.
DEAG advised that any future approach should clearly define the purpose, limits and safeguards for data collection and use. Members supported principles such as collecting only what is necessary, using de-identified information where possible, ensuring data is ‘as open as possible and as closed as necessary’, and providing clear justification for decisions and actions.
Members also discussed the need to build and maintain social licence. This includes meaningful public consultation, informed consent, transparent communication about benefits and risks, and proactive guardrails to prevent function creep, such as use of the technology for purposes beyond safety.
Māori data considerations and personal data sovereignty were identified as important areas for further work. Members advised that the technology should not be mandatory in the New Zealand context, while noting that broad public participation might be needed to realise safety benefits.
DEAG recommended beginning with a ‘worst case scenario’ assessment to identify what could go wrong, then developing legal, policy and technical safeguards to prevent unacceptable outcomes. Members also suggested learning from existing safeguards in other data contexts, including the governance of secondary use of administrative data.
The presenter noted there was no government commitment to progress the technology at the time. DEAG offered further support to help NZTA consider how connected vehicle technology could be explored in a way that protects privacy, supports public trust and enables innovation with confidence.
The meeting ended with a karakia.
Members: Professor Colin Simpson (Chair), Jonathan Godfrey, Jonathan Kilgour, Kate O’Connor, Russell Craig, Will Koning
Apologies: Andrew Sporle
Presenters: Stats NZ representatives
Meeting opening and standing items
The Chair noted the apologies.
CDEI noted that work, which started last year, had focused on creating a data ethics framework centred on human values to enhance ethical decision-making, act as a connecter to other established frameworks (e.g., Ngā Tikanga Paihere), and to provide accessible upskilling, engagement and transparency. Presenters sought DEAG’s feedback on the framework before presenting it to the Stats NZ Strategic Governance Board for their approval in late May.
Presenters noted some data ethics harms:
The DEAG members reflected on the Data Futures Work NZ and its current relevance and usability. They discussed how this framework could stay relevant long-term, being an ongoing conversation, rather than one and done.
CDEI provided an overview of the HVDE framework on how data collection and use might impact individuals and communities, helping organisations to pause, question and discuss in order to surface ethical impacts early and make them more visible. They noted the framework was focused on developing guidance shaped by a set of human values, designed to be wide rather than deep, to enable uplift to ethical capability, removal of silos and ethical use of data.
CDEI presented the core six values that uphold the framework (including what they encapsulate), which cross- reference to other guidance (e.g., DPUP, Ngā Tikanga Paihere) and each other:
CDEI and DEAG discussed that values could be toggled, but making clear when they are applied with careful judgement. DEAG advised being practical about values in different contexts as some might have more weight than others.
DEAG supported the framework, raising the need for a balanced approach to honouring individual and collective worldviews to avoid weaponizing of processes.
CDEI noted that some practical steps to introduce the HVDE framework had already been taken at Stats NZ, however, a structural approach to rolling out the framework would still be needed to create a baseline of ethical literacy by:
DEAG advised asking if legislation or government policy required to be unethical, noting the framework would structurally address issues for targeted communities with importance of involving their authorities. DEAG noted a learning system would be needed to address gaps and enable feedback loops, as well as support from the data ethics authorities. DEAG advised that the framework was ready to be implemented, suggesting inclusion of case studies and looking at relevant jurisdictions.
The meeting ended with a karakia.
Advisory Group Members: Professor Colin Simpson (Chair), Andrew Sporle, Russell Craig, Kate O’Connor, Will Koning, Jonathan Kilgour
Apologies: Jonathan Godfrey
Attendees: Social Investment Agency, Stats NZ
Facilitators: Stats NZ representatives
The Centre for Data Ethics and Innovation facilitated a workshop with the Data Ethics Advisory Group (DEAG) and the Social Investment Agency (SIA) to explore ethical considerations across the social investment approach. Participants discussed the importance of doing good while doing no harm, how to build genuine partnerships with communities, support informed consent, and ensure people understand how their data will be used. The group highlighted the importance of trust, transparency, and social licence throughout commissioning, data collection, analysis, and use of insights.
Across the stages of the social investment approach, participants identified areas where ethical guidance and support would be valuable. These included improving transparency about data use and quality, addressing potential bias, and ensuring communities are involved in shaping measurements and interpreting findings. The workshop also highlighted the value of independent peer review and evaluation to strengthen trust and quality.
The group emphasised a need for learning systems that support safe, transparent, and adaptable approaches. This included involving service users in oversight, publishing insights safely, creating feedback loops to improve practice over time, and driving system change. DEAG members discussed ways they could support this work, including linking SIA with experts to assist with peer review and evaluation.
Members: Professor Colin Simpson (Chair), Andrew Sporle, Jonathan Godfrey, Jonathan Kilgour, Russell Craig, Will Koning
Apologies: Kate O’Connor
Attendees: Centre for Data Ethics and Innovation (CDEI), Stats NZ
Presenters: Topic One: Ministry for Ethnic Communities
Meeting opening and standing items
The Chair noted the apologies.
Presenters clarified the definition of the population they focussed their project on, comprising Middle Eastern, Latin American, African+ (MELAA+), including population missing from the Stats NZ’s definition of MELAA (Middle Eastern, Latin American, and African).
They noted the increased population coverage is used to provide more meaningful insights to sample surveys and follows the Ethnic Communities Ethnicity Classification.
The research under discussion supports the Ministry for Ethnic Communities (MEC) in their commitments under the 2025–29 Suicide Prevention Plan. Presenters also noted that MELAA are two percent of the total population, and this study focussed on 57,000 children and young people under 25 years old, with the majority of them living in Auckland, and the largest group being of African ethnicity.
The DEAG members were advised that the concern around the mental health of the MELAA youth population stemmed from the ethnic communities themselves and the Ministry of Social Development’s What About Me? survey of youth from 2021, which reported concerning MELAA youth mental health wellbeing including high suicide ideation levels. Because of the small numbers surveyed, the sample error on these estimates was large. An admin data solution sought to provide a firmer evidence base.
Presenters outlined the mixed method design approach used by MEC:
The DEAG members discussed the admin data view comprising data from secondary, special mental health services.
DEAG queried if data for youth in the care of Oranga Tamariki or Department of Corrections would be included in the dataset (especially health-related) and was advised that Oranga Tamariki’s jurisdiction and records of ethnicity were considered by MEC. Members noted the MELAA group was a small population, and that any improvements in Oranga Tamariki’s collecting, mental health recording and ethnicity system could also be used for other populations.
DEAG noted the critical outcome of the project would be the identified gaps in data, while highlighting the importance of learning systems and understanding the data ecosystem. They added that rather than relying on surveys, MEC should focus on routinely collected data; presenters in return advised that a code module had been developed by MEC to identify IDI data relating to Ethnic Communities’ populations.
The group discussed what gaps exist in the IDI and what data are not visible (e.g., GP data, private services such as faith-based alternatives), advising to use the IDI Search tool to double check. A member advised to consider whose data matters (e.g., asking wider MELAA population not just youth using service) and raised concern around starting research with admin data due to its limitations. They suggested that the wider young MELAA population should be asked directly, rather than relying solely on youth accessing health services. They noted the risk of re-traumatisation during engagement and were advised that MEC were looking into the Ministry of Health Youth Advisory Group’s work to ensure young people feel safe and protected during the process.
The DEAG members suggested looking into identified risks (e.g., welfare, pastoral, justice and health) from the work done by Education NZ on a whole-of-system value and cost for international students, using IDI and primary research.
DEAG advised presenters to seek more granular data on refugees and liaise with MBIE and the Auckland Refugee Centre.
DEAG were supportive of the project, its future significance and suggested to MEC to:
Presenters acknowledged the risks of re-traumatising young people and findings being statistically insignificant; DEAG advised ensuring information is robustly collected and focus on the importance of the work.
The minutes from the meeting on 11 December 2025 were approved as a true and accurate record of the meeting.
Members noted the action register.
Members discussed potential future agenda items.
The meeting ended with a karakia.