We will never ask you for your data: Māori GIS, control and the difference between access and authority
One sentence eventually became one of the clearest principles in Ngā Poutama Matawhenua: we will never ask you for your data. It sounds like a small operational rule for a training programme. In reality it contains much of what twenty years of Māori GIS work taught me about authority, trust, rangatiratanga and the limits of technical access.
GIS people are trained to acquire data. We search for layers, download services, request files, convert formats and build databases. The professional instinct is often that more information is better. In many ordinary projects that instinct is useful. In Māori GIS it can become dangerous if the technical desire to collect information gets ahead of the questions that should come first: whose information is this, what relationships does it sit within, who has authority over it, why is it being collected and what responsibilities come with holding it?
I am not Māori and do not claim to define tikanga for other people's information. What I can describe is the change in my own practice after years of working with claimant information, iwi and hapū projects, Māori GIS practitioners and training programmes. The more I learnt, the less comfortable I became with the ordinary technology assumption that the person who possesses the file is therefore entitled to organise, copy, analyse and publish it. Technical custody and legitimate authority are not the same thing.
That distinction sits close to rangatiratanga. In a GIS context, I have come to think about it practically: who can decide what is collected, who can see it, where it lives, what leaves the organisation and whether a map should be made public at all? A platform can give people access while quietly shifting those decisions elsewhere. A community can be invited into a modern digital system while losing control over where the information goes. The interface may look empowering while the architecture does the opposite.
Getting the file
In the early claimant mapping years, a great deal of effort went into finding information. Historical plans were scattered across research files, archives and government systems. Modern cadastral data had to be sourced. Imagery could be expensive or difficult to obtain. Place names appeared in multiple forms. Claimant researchers had their own records. Some information existed only on paper or in people's memories.
The GIS challenge was to bring useful material together. That was powerful because spatial relationships could suddenly be seen. A historic block could be compared with a present road. A Public Works taking could be viewed against current property boundaries. Historical names could be related to modern geography. Researchers could move between documents and maps more easily.
But bringing information together introduced another problem. The GIS team could end up holding copies of material supplied under very different circumstances. A historical map might have been lent for one purpose. A set of significant places might have been mapped for a claimant process. Research notes could contain material intended only for a defined audience. From the computer's point of view these were all files. From the people's point of view they were not interchangeable.
That difference is fundamental. Information can carry relationships and obligations that do not appear in the file format. The shapefile does not tell you who trusted whom when it was created. The database does not tell you whether the kōrero behind a location was intended only for a small group. The cloud service does not know whether the person uploading the information actually has authority to do so.
Technology tends to strip information from context unless people deliberately put that context back.
Whakapapa of information
One useful influence on how I now think about data is whakapapa, although I use the concept carefully. A GIS metadata record is not whakapapa, and a database cannot reduce whakapapa to a lineage field. But the idea that information comes from somewhere, is connected to people and place, and has a history of transmission is much richer than the usual technical assumption that data is simply a reusable object.
Every important layer should have a story of origin. Who created it? Who contributed to it? What was its original purpose? What changes have been made? What authority did the contributors have? What conditions came with its use? If it has been generalised or derived from something more sensitive, what relationship should remain visible between the public version and the source?
This is ordinary provenance in one sense, but the Māori GIS context pushed me to treat provenance as more than a quality-control field. It helps prevent information from being detached from the relationships that gave it meaning. Once a layer loses its history, the next person tends to treat it as generic data. That is where accidental misuse begins.
A folder can survive longer than the people who created it. Ten years later someone opens the GIS and sees a layer with a convincing name. Unless the history travels with it, the geometry may outlive the authority that justified it.
Possession is not permission
Digital systems make copying so easy that it can almost disappear as an action. A person can duplicate a layer, add it to another project, put it in a shared folder or publish it online in seconds. There is no physical friction to remind them that the information may have arrived through a relationship, agreement or particular purpose.
Possession is therefore a poor proxy for permission. If someone gives a GIS practitioner a sensitive layer so a specific map can be prepared, that does not automatically mean the practitioner can reuse it elsewhere. If a claimant group supplies information to support a hearing, that does not make the information a general resource for anyone who later encounters the file. If a map is borrowed and digitised, digitising it does not transfer rights over the knowledge represented on it.
This seems obvious when stated plainly, but software rarely enforces the distinction. A shapefile has no moral memory. It does not know why it was created, who trusted whom, whether a location is sensitive or whether the file was intended for one project or many. Those responsibilities remain with people.
That is where I see a practical connection with kaitiakitanga. I do not mean that data stewardship and kaitiakitanga are interchangeable terms. They are not. What influenced me was the idea that holding something can create responsibilities that extend beyond today's task. If information matters to a community, the question is not simply how efficiently it can be stored. It is how it should be cared for, who should have access, how its context is protected and what must be passed on to those who inherit responsibility for it.
Not everything wants to be open
I am a strong supporter of open public data where the information is genuinely intended to be public. Better access to LINZ, council, environmental, Census and other government spatial information has transformed what communities can do. Data that once required personal contacts, purchase or specialist systems can now be accessed directly through web services or download portals.
The problem is when the open-data instinct becomes universal. There is a tendency in technology to assume that information becomes more valuable as it becomes more open. That may be true for many public infrastructure datasets, elevation models, roads and environmental basemaps. It is not automatically true for information connected to whakapapa, places of significance, local histories or culturally sensitive interests.
Some information gains value through appropriate restriction rather than maximum circulation. A community may decide that a generalised public layer is suitable while detailed source information remains local. A place may be shown approximately. A name may be public while associated kōrero remains restricted. Some information may be visible only to defined users. Some may not be appropriate to digitise at all.
There is nothing technologically backward about that. It is a deliberate expression of authority.
The purpose of GIS is not to maximise the number of layers on the internet. The purpose is to support the kaupapa for which the information is being used.
Authority before architecture
Technology projects often choose the platform first. The organisation decides it wants a cloud GIS, a data warehouse, a web application or another system and then starts working out how the information will fit inside it. For Māori GIS, I increasingly think the sequence should be reversed.
Start with kaupapa, authority and relationships. What is the work trying to achieve? What information is involved? Who has mana whenua or other legitimate authority in relation to it? Who needs to use it? What can be public? What should remain controlled? How long does it need to exist? Who will carry responsibility when current staff move on?
Only then should the architecture be chosen.
This matters because modern systems are very good at encouraging centralisation. Platforms want to store the data, manage the users, publish the maps and become the place where every workflow happens. From a technical perspective this can be efficient. From a rangatiratanga perspective it can also create dependencies that are hard to unwind later.
A community-controlled GIS does not necessarily need to be large or elaborate. In some cases a carefully managed QGIS project, local files, a backup process and selected public services may provide more genuine control than a sophisticated cloud environment. In other cases a cloud system may be entirely appropriate. The right answer depends on kaupapa, authority and capability, not fashion.
Training without taking
This was the thinking behind the Ngā Poutama principle that we would not ask participants for their data. Training programmes often request real datasets because they make exercises feel more relevant. There is some logic to that, but the request can create an unnecessary transfer of control. The trainer may suddenly possess material that was never intended to leave the participant's organisation. A workshop platform may receive information nobody has assessed for sensitivity. A temporary account can become a permanent copy somewhere in the cloud.
Most technical teaching does not require that transfer. If I am teaching how to build a point layer, configure symbology, set up a form, georeference imagery or organise a StoryMap, I do not need a participant's actual culturally sensitive information. A sample dataset can demonstrate the method. The participant can then repeat the process locally with information they control.
That approach also changes the relationship between trainer and participant. The trainer is there to increase capability, not to become the custodian of the participant's information. The participant does not need to surrender data to prove that the training matters.
There is an element of manaakitanga in that approach too. Good support should reduce unnecessary risk for the person receiving it. It should respect the fact that they may be carrying information on behalf of whānau, hapū or iwi and may not have the authority to hand it to an outside trainer simply because the software exercise would be easier.
Public, shared and restricted
One of the practical challenges for any iwi or hapū GIS is that the same broader system may contain information intended for very different audiences. A public-facing map might show marae, waterways, environmental projects, publicly recognised place names and selected historical information. An internal system may contain operational or planning material. A more restricted environment may contain culturally sensitive information, personal information or locations that should not be exposed outside a defined group.
The mistake is to assume one technical platform automatically suits all of these. A useful design separates authoritative source information from presentation layers. Public StoryMaps or web maps can use approved and generalised information without requiring the complete source dataset to become public. Internal operational layers can sit behind authentication. Sensitive material can remain outside the public publishing environment altogether.
The public map is therefore a view, not necessarily the source. That distinction protects both utility and authority. It allows a community to communicate widely without assuming that communication requires exposing everything it knows.
This is one area where Māori concepts can challenge the default settings of technology. Western information systems often treat visibility as a sign of value and openness as the preferred end state. A Māori framing can ask a different question: what is the right relationship between this information, these people and this purpose? Sometimes the right relationship is public. Sometimes it is shared within a defined group. Sometimes it is deliberately kept close.
Rangatiratanga and choice
Control is meaningful only when people have choices. A system that claims to support local control but cannot export the data cleanly is not offering much choice. A platform controlled by one consultant's personal account is not strong local control. A community that can publish a map but cannot decide where the source information is stored still has a dependency.
This is why technical design details matter. Open formats, clear account ownership, backups, documented permissions, local copies and exit plans can all support rangatiratanga in practical ways. They are not Māori concepts in themselves, but they can be used in service of Māori authority rather than against it.
The same applies to software choice. QGIS can be useful because it removes one licensing barrier and supports local files, but open source does not automatically make a project culturally appropriate. ArcGIS can be appropriate where the organisation has the capability and governance to use it well. Cloud systems can support collaboration. Local systems can support greater isolation from external services. There is no one culturally correct software product.
The important question is whether the technology serves the kaupapa and leaves meaningful decisions with the people who should make them.
AI changes the question
Artificial intelligence makes these issues more urgent because the boundary between using a tool and giving the tool information can be easy to overlook. A person pastes text into an AI assistant to summarise it, uploads a spreadsheet for analysis or provides coordinates and asks the system to generate a workflow. The interaction feels conversational, which can make the information transfer feel less formal than uploading data into a database.
The information has still left the local environment.
For Māori GIS this requires the same discipline that should apply to cloud systems. What information is being provided? Is it public? Is the user authorised to upload it? Does the system retain it? Is the service appropriate for culturally sensitive, personal or controlled information? Does the convenience of the tool quietly shift authority somewhere else?
AI can be extremely useful without being given everything. Public datasets, synthetic examples and de-identified information can be used to generate workflows, scripts and teaching material. An AI system can explain how to build a QGIS expression without receiving a community's sensitive layer. It can help design a schema without receiving the real records.
The old training principle applies perfectly: teach the method without taking the data.
The right to leave things unmapped
One of the strongest lessons from Māori GIS is that capability includes the ability to decide not to map something. Digital technology often presents mapping as an obvious improvement. If a place can be recorded, why not record it? If a layer can be published, why not publish it?
Because the people with authority may decide that the information should remain in another form. Some knowledge may sit more appropriately in kōrero, in relationships, in local practice or within restricted records. A coordinate is not inherently a better way of holding knowledge. It is simply one representation.
This is where the concept of taonga also requires care. Some information and knowledge may be regarded as taonga, but a GIS practitioner should not simply label every dataset a taonga and assume that solves the governance problem. The practical implication is more demanding: if information is precious, relational or culturally significant, the people connected to it should have a meaningful say in how it is represented, copied and shared.
The right to map and the right not to map are both parts of genuine control.
Generalisation as a respectful choice
One of the most useful techniques in sensitive mapping is generalisation. The choice is often presented as either publish the exact data or publish nothing. There are many options between those extremes. A detailed point may become a broader area. A sensitive feature may be represented at catchment scale rather than by exact coordinates. Public maps can omit attributes unnecessary for the story. A detailed internal layer can produce a simplified public derivative.
This should be done deliberately and with the approval of the people responsible for the information. It is not a technical trick for bypassing authority. It is a way of matching the level of disclosure to the kaupapa.
The better question is not simply “Can we put this on the map?” It is “What is the minimum information needed for this purpose, and who should decide?”
That is a much better design question.
Control is capability
Over time I have come to see data control as part of capability rather than a separate governance concern. A group is not fully capable if it can make maps but cannot move its data. It is not fully capable if one vendor account controls access. It is not fully capable if nobody knows which layers are appropriate to publish. It is not fully capable if the system requires sensitive information to leave the organisation unnecessarily.
Capability means being able to make informed choices. Sometimes the informed choice is to share widely. Sometimes it is to publish a generalised layer. Sometimes it is to keep information local. Sometimes it is to decide that the information should not be digitised. The GIS should support those decisions rather than silently make them on behalf of the community.
That is why the Ngā Poutama sentence matters to me. “We will never ask you for your data” is not an anti-technology statement. It is a pro-rangatiratanga and pro-capability statement. It says the purpose of the training is to increase what people can do without making participation dependent on surrendering information they hold on behalf of others.
After twenty years around Māori GIS, that feels like one of the better tests of whether technology is serving the kaupapa or the kaupapa is being reshaped to suit the technology.
For the wider history, see From DVDs to AI. For the related article on long-term capability, see After the mapper leaves.