Introducing the Framework for the Governance of Indigenous Data

Good morning everybody. Uh it's so terrific to be here with you again today uh to talk about this exciting new development the publication of the framework for indigenous data governance and two of the I'm going to acknowledge the traditional owners first of course I haven't got my PowerPoint up So it doesn't it doesn't go in there. >> Yeah, why don't I do it there? Thank you. Thank you. Clarity. Uh we acknowledge the Garner people, the Aboriginal people uh traditional owners of the lands where we meet. We pay our respects to their elders, past, present, and emerging and the place of indigenous knowledge in the academy. So, uh, I went looking for some images of the Ghana people last night and found, uh, this incredible painting by Alexander ShraMM from 1850, um, Adelaide, a tribe of natives on the banks of the river Torrance. Um, and you can see, uh, how many people there were. And, uh, you know, it's a village. and and soon after that they were moved out of the area. But the next image is a map of the Ghana people's native native title determination area. So it's um a small act of justice for the uh terrible impacts of colonization on the Ghana. So I want to be very clear that we appreciate being on their country and we acknowledge their traditional ownership. So, just a little bit about our work. Well, who we are and nothing works. It's not moving. Another arrow. Different arrow. That one. There we go. Right. So the Indigenous Data Network was founded in late 2017 following the inaugural Australian Indigenous Data Sovereignty Symposium held at the University of Melbourne which Kristen and I convened with others from the Australian Institute of Aberinal and Toouristra Islander Studies and the ANU. So this is a national network and we aim to provide a collective base from which to support and drive better data outcomes for ainal and toristra islander peoples. The network aims to support indigenous communities. Well, we don't just aim to, we do uh and community controlled organizations to collect, manage, and utilize data assets relevant to their needs and aspirations and promote indigenous data empowerment. So our uh project funded by the Australian Research Data Commons um well there's actually two of them and I don't want to take up a lot of time here because uh the project that Robert is involved in um also funded by the ARDC Alaka which you heard a bit about yesterday but not enough. So just to say our project is uh titled improving indigenous research capabilities building an aberinal and toouristrate islander data commons and you'll remember that I put the the website up yesterday. So the purpose of that of our project is better understanding of indigenous research communities across Australia, better understanding of indigenous data across Australia, development of an indigenous research data catalog and we've done that and Les Nebones there he's been working on that with um Nick Carr, they'll talk about that later. um data rematriation prototyping. So right we're girls right? So instead of repatriation we think rematriation is a better way of describing the process. Um technical architecture based on cat cataloges of data assets using DCAT 2 international standard for interoperability between data cataloges. And as I said, you'll hear about that later. Training and education to build indigenous research data cap capabilities across the nation. So, so far we have one accredited indigenous data governance subject. But you know events like this also I believe provide foundational training in these matters. Um so we have substreams of work and teams of people working on the substreams uh indigenous data catalog resources and extensions indigenous spacio temporal frameworks and infrastructure. So if you go to our indigenous data commons site, you'll see an interactive map um indigenous uh data capability building for digital futures, indigenous repositories and data rematriation. Um and these are the main elements of what we regard as indigenous data governance and sovereignty. Um we have five communities uh in our network. Um so there are five communities of indigenous research data custodians. There are organizations themselves, universities, researchers and other research bodies. You remember yesterday I talked about orphan data sets as one of the problems uh for that community. um the glam sector, galleries, libraries and museums um and you know archives um Commonwealth State and Territory and local governments and the private sector. Um so that that's a list of all of our partners and our project team. Um so let me hand over to Robert on Aldaka. Thank Thank you, Marcia. Um I'm going to juggle some laptops here. Doesn't fall off. Okay. So, I'm just going to speak a little bit about um Eldaka, which you would have heard. Um we took a little deep dive into the Arocorrate method yesterday with my colleagues Simon and Ben. Uh and that was that was very good as well. So you already had a little sort of taste but we haven't given you the general kind of overview. So I'll try and do that somewhat briefly. Um this is a slide that really talks about what we're doing. LDAC is the language data commons of Australia and I guess I want to talk a little bit about why why we're established and why we're doing the work that we're doing. There are more than 400 languages actively spoken in Australia and of course on this continent it's home to the oldest continuous living cultures um oldest continuous living civilizations in the world and also to Australia is situated in one of the most uh linguistically diverse regions on the planet that being the Asia-Pacific region. So there's a lot of languages spoken in this area of the globe. And also too, one of the other reasons why we're doing the work that we do is the fact that generative AI and machine learning are now fundamentally changing the ways uh in which researchers and communities can engage uh with language and text uh textual data. It's dawned on me that I have missed uh our partners, our collaborators slides. So, I really do apologize for that. Um but we we have a uh I'm going to try to do it off the fly now and embarrass myself. We have a few partners. We have um uh we collaborate with the A&U, Sydney University, Melbourne, University of Melbourne, the Bachelor Institute of Tertiary Instit uh tertiary education, uh the AIDC, uh the uh First Languages Australia, which is the peak body for indigenous languages in Australia. >> Simon, have I miss I've got everyone. UWA. Yes. Yeah. Okay. So, this slide I want to share with you is just looking a little bit around the structure, how we kind of conceptualize data. Specifically, we're working with language and cultural data. So often a lot of the materials that you will find that are in glam collections that are in university repositories uh in an indigenous data context it is uh a lot of that material uh language manuscripts could be old recordings on wax cylinders. it. The format of the data is really diverse and it's often in precarious or uh uh situations that aren't so secure like the example we gave the other day of uh m you know manuscripts in boxes in the back shed you know so they're not necessarily safe uh safely kept some are but also to the work of Aldaka does extend slightly beyond uh indigenous data to encompass uh other other language material aterials like old English, um, uh, migrant and minority languages, uh, uh, oslang sign, that sort of stuff as well. But when it comes to language and cultural data for us, uh, you'll see up there at the top slide, it's about it's about finding, securing, collecting, and organizing. Now, when we look at the tools that we use and particularly the skills that we need, it's about analyzing the data. It's about repurposing it. And it's about training those who work with the data. So that's a big part of what we do also. Uh and always at the center of that it is people data. It's data about us. Uh and in some instances it's data by us. But at the heart of the project we seek to contribute to indigenous data ecosystems with a strong emphasis on data for us. um whereby indigenous communities can truly derive benefit from data in meaningful ways that center community priorities um community determined priorities and I think that was a big driver for why we had uh such an active involvement in the framework because that's what it seeks to do and I will these are some of the challenges that we have faced uh particularly in regards to the condition of language and cultural materials. I'm not going to go through them, but please feel free to take a photo. Um, and certainly this is not conclusive by any means, but uh what we found thus far. So, please if there's other ones, I I'm very interested in hearing your um data stories or else the condition of of them as well. Uh the other one I would like to share with you just in closing, actually I got two more. This is our activity streams. This is how we're structured across uh five different activity streams. Um this is the way in which we work and then we're broken down into like 101 different work packages. So there's a a lot of that's the way we're structured. That's the way we're going about uh managing the work that we do because a lot of people will say well where do you start? How do you it's it's a mammoth task. Even even one priority one priority of trying to find all indigenous language material that exists is uh seems insurmountable but it's an ambitious goal one of many that we have and the last one actually sorry second last one I want to share with you um this the the purpose of this slide is just to really give you uh a demonstration of where some of the that indigenous data in here particularly language and cultural data. So that's some of the places in which we find it. Uh I'll probably finish there but this one is just to say that um as as was made clear in Ben and Simon's pres uh presentation yesterday that the there's a strong importance on the long-term preservability of data. Uh and the this just echoes that um we have the pillars pillars protocols there which you can uh scan that QR code and you'll you'll get to see um the pillars protocols that we developed for implementing long-term archival repositories. Uh and also to this one is just highlighting the Arocrate standard and packaging the metadata uh with the data all together through that process. Uh, I will hand my hand back to you, Marcia. So, this is more or less an official launch, isn't it? Uh, so three of the authors are here, Jenny's over there, and there's somebody missing. Who's missing? That's it. It's us four. Yeah. So, uh this is the framework for the for the governance of indigenous data, humanities, arts, social sciences and indigenous research data commons. Um it's downloadable. Um if you go to the ARDC website and find >> it's >> and it's in your welcome pack as well. Yeah. So, there are no printed copies. >> I thought it was because we were so green, but anyway, uh you're proposing to publish some, are you? >> Yeah. >> Yes. So, that's a very good point of view. Um so the document is not a static document. It will change over time and so publishing hard copies you know would be a a waste of resources given that uh this will develop very quickly especially as people like yourselves begin to use it. So uh importantly here and I spoke about this yesterday in this document we have a definition of indigenous data data generated intentionally or not by about or for Aboriginal and toouristr islander people. Indigenous data refers to information in any format or medium collected, analyzed, stored, and interpreted within the context of indigenous individuals, collectives, populations, entities, lifeways, cultures, knowledge systems, lands, biodiversity, water, and other resources. It includes data collected, used or stored by any agency, department, laboratory, organization, corporation, statutory body, university or research institute conducted by, with and about indigenous people or that should say or peoples. I hope it's not a typo in the book. Um, and data that indigenous communities have generated and maintained themselves. So you can see that we've been very comprehensive because and I'm not going to say that you know this is the most fought over contested definition in the history of the Australian government funding indigenous efforts but it's got to be close to it. Um people who um are concerned with fish and oceans objected to this in the first instance. Um and you can see why you know they want to exclude us from their their data sets. So this was developed some years ago as I said yesterday at roundts like this where people raise their concerns about what could indigenous data sovereignty mean in practice. Well unless we have a strong comprehensive definition like this you know we're talking word salad. So uh I wanted to particularly draw your attention to that and you'll note that yesterday and today we'll talk about indigenous data assets and that's a different category but just so you know that we have put a lot of work um into in into the definition. um and it rem there's work work that remains. We need to define other concepts as well. So we've done some thinking about indigenous data sovereignty as a concept. Sorry. Um and I want to make this point here briefly today. The concept of indigenous data sovereignty has emerged as a powerful manifesto of indigenous people's rights to control data that relates to their communities, cultures, lands, and knowledges. It is grounded in the broader political and cultural struggle for self-determination and the reclamation of authority over indigenous knowledge systems. And you know it was really Cucatai and Taylor who brought this concept uh to our attention in their publication in 2016 which is downloadable from the A&U press website by the way and there's been a lot of literature since in Australia the movement has gained momentum in response to the historical and ongoing extraction misrepresentation uh and manage mismanagement ment of indigenous data by state and institutional actors. Although the language of of sovereignty may hold political resonance, it is also conceptually and legally fraught in contexts such as Australia. Indigenous sovereignty is not formally recognized by the state and as such lacks enforcable legal status. This creates a fundamental tension. Indigenous communities may assert sovereignty over data but without institutional sorry without institutional mechanisms to uphold these claims sovereignty remains largely symbolic. The concept of sovereignty traditionally associated with supreme authority, independence and self-governance within defined territorial boundaries and of course it originates in Europe um in the context of wars, conflicts and so on. um has been a foundation of international law and political theory again most of it European in the Westfallian model of nation states sovereignty encompasses territorial integrity legal authority and non-intervention principles however applying this concept to indigenous contexts particularly when addressing indigenous data presents significant theoretical and practical challenges So we've written about and it's not published yet several of the epistemological and ontological obstacles. The digital nature of much data transcends geographical boundaries complicating jurisdictional delineations and that's what most of the big debates in the world are about. crossber flows. That's what the debates are about. And we know now with the, you know, AI seeping into every aspect of our world, talking about borders and sovereignty is not a winning strategy. We need to come up with a some more ideas. Um the frequent traversal or flow of data across multiple jur legal jurisdictions problematizes the assertion of exclusive sovereign control sovereign control. This complexity is a particular is particularly acute for indigenous peoples around the world whose existence within nation states that claim overarching sovereignty creates potential conflicts in the assertion of data rights. The intangible and fluid nature of data further challenges conventional notions of possession and control central to sovereignty discourse. Finally, and most importantly, in Australia, data are legally held by government authorized data custodians under Australian law and hence these data custodians have substantial ownership and control rights over the data. The possibility of indigenous peoples or entities successfully claiming ownership and control over such data is extremely unlikely. At present, the DAT act as I talked about yesterday does not include any mention of indigenous people, indigenous data or data relating to indigenous people. So, you know, I won't uh go on any further, but just to say that, you know, to sum up the rest of the argument here, data is always relational, right? We could too easily be dragged into you know cooker theories of um what's their mad idea sovereign citizen sovereign citizen right I think a lot of the cooker theories of the sovereign citizen have infected our world and we you know we have to be very clear we have to be very scientific and rational about this um you know for those of you don't know what I'm referring to there. Desi Freeman, right? That kind of sovereign citizen. Um, so we say that the point is to negotiate these rights under the principle of self-determination. If we're going to talk about something meaningful, we need to operationalize self-determination and negotiate uh the rights. And that means of course, you know, very strongly asserting on behalf of our espe especially our community control sector, their ownership of their indigenous data assets. When we first started talking to Canberra about this, they, you know, literally had no idea what we were talking about. Um, I can't get this to work. >> Oh, the right arrow. Yeah, I'm pressing the white. I'm pressing it. Oh, I've I've pressed all of them. None of them are working. It's fro. Oh, there we go. Indigenous data governance. And the next topic is indigenous cultural and intellectual property. So Kristen, do you want to take over at this port point and talk about indigenous data governance as the alternative and the only rational alternative to the concept of data sovereignty? >> Sure. Thanks, Marcia. I'm not going to go into it in huge detail because these are all concepts that you can read about in depth in our framework. Um, and it it it really just sets out some of the rationale that our projects are are drawing on and hopefully it sort of clarifies some of these issues. But we really are using indigenous data governance um in general rather than indigenous data sovereignty. Um, our our logic is that we think in in Australia it offers a more pragmatic and culturally appropriate framework for addressing the complex challenges of data man management in indigenous contexts. And this is sort of the way that we're thinking about it in the context of um research data um in our day-to-day working lives when we work with different communities, different data sets, uh different structures, with government authorities. Uh it just seems to provide a far more um tangible way of approaching these issues related to power and data in Australia for indigenous people. Uh we also have a section in the framework where we set out um out how we refer to the rights of um Aboriginal and tourist rate islander peoples in controlling, protecting and benefiting from their cultural heritage. So indigenous cultural and intellectual property. Um, article 31 of the United Nations Declaration on the Rights of Indigenous Peoples affirms these rights encompassing cultural expressions, oral traditions, ecological knowledge, and intellectual property derived from indigenous sciences, technologies, and practices. And in the work that we're doing, we're finding regularly that this is a a growing issue for indigenous people across Australia and undoubtedly in internationally. In Australia there there aren't any um legal frameworks that actually protect ICIP. Obviously we have copyright laws which do protect some of this but there's huge comp complexity when we start thinking through the problems that are faced um by all of us on a day-to-day basis. For example, I was up in um Anamland in Yala last week working um we've got a one of our activities where we're working with the bilingual school up there. They've got an archive of about 50 years or more of bilingual material and trying to untangle the copyright, the IP and the ICIP in just one archive is extraordinarily complex. Um, it requires a great deal of of legal expertise, but also even trying to understand who should be the legal holder of these rights. So, it it's not a simple issue and it's one of the big challenges I think that we need to face across Australia. But as I said, we've got a section in the in the framework document that goes into this in a little more detail. But just to introduce um for those of you who haven't had a chance to read through the framework document yet, the governance model that we're proposing. So it's built around five components or guidelines. The first one being recognizing indigenous data assets. So that's acknowledging the ownership and rights of indigenous people and entities over their data including observing their legal rights to use, share, store and report such data and complying with their decisions while also supporting their status as data custodians. Uh the second guideline is to partner with Aboriginal and tourist rate islander people. So this is to make sure that um there is collective benefit, meaningful representation of and collaboration with indigenous communities, respecting their perspectives and complying with their rights. Um the third one is building and enabling data related capabilities. So this guideline is about empowering indigenous data custodians and stakeholders with the skills and resources necessary for effective data management and utilization and especially data ecosystem maturity including cyber security measures to enable their qualification for data access. Uh the fourth uh capability is looking at self self-determination. Uh so that's providing knowledge, accessibility and usability of data assets. Making sure that indigenous communities and and of origin and peoples have access to and benefit from data resources. Emphasizing transparency, inclusive inclusivity to galvanize and support indigenous self-determination. Uh the fifth guideline is is looking at building an inclusive data ecosystem. So making sure that we're fostering collaboration and inclusivity within the research data landscape to promote and foster greater accessibility, collective growth and understanding. So as you go through and have a look at our framework, we go into a great deal of detail about how you as data custodians or researchers or archavists can actually look at implementing or operationalizing this framework. We draw on a number of principles including the fair and care principles predominantly the care principles which you would have heard a lot about yesterday as well. Uh in terms of recognition of of indigenous art assets, as Marcia mentioned yesterday, a lot of government agencies who uh provide funding or or pol write the policies uh have very little understanding of indigenous assets held by indigenous people across Australia. Whether we're thinking about the hundreds of ACOs across Australia who hold medical and and other data for different community members or whether we're talking about the legal agencies, whether we're talking about a variety of ACOs who develop and gather their own data. Um so we thought it was very important as I said whether it's government agencies, universities or institutions to recognize these assets and the the vital detail that's held in these assets. So in terms of the governance logic of this this particular guideline, recognition affirms the legitimacy of indigenous governance systems and the authority of indigenous communities to determine how data about them is collected, interpreted, and used. It challenges colonial data practices that have historically excluded indigenous voices and repositions indigenous people as central agents in data governance. and its conceptual role is in the foundation of of the framework. So this recognition underpins and enables all of the other governance um functions within the HS and indigenous research data commons. It legitimizes indigenous authority and sets the conditions for ethical research engagement. Do you want to jump up? Thanks. Uh so we're moving on to partnerships. Did it again there. So I want to make uh this clear. What what we've discovered in our work is that a formal contractual partnership uh is a huge benefit um and is very clear whereas a a non-contractual collaboration um you know leaves too much uh room for doubt and it's not a great trust building platform. So if we can transition what are you know very often very happy collaborations into a formal partnership it protects everybody's rights. We just to say like this is an aside we're we're pretty concerned about how universities uh assert intellectual property over other people's um intellectual property especially cultural property because they don't know what it is. I can't imagine that there's cultural property in it as an out, you know, as an input to a research project. They just can't see it. Too long a story to explain, I suppose. But the only thing they will recognize are the intellectual property uh rights of academics who publish. And you'll notice that a lot of researchers now are are co-authoring with the people the ainal people that they've been conducting research with. So that's a good simple measure, but we have to think of more ways to protect people's intellectual and cultural property rights in research outputs um and in archives. So this problem that Kristen's wrangling in Arnamland really difficult you know uh but it can be done um it's we we will probably have to have a workshop for instance and I'm trying to make this very practical for you'll have to have a workshop people like us in this room on um intellectual property disclosures and registers you know what terminology can we develop to best register indigenous intellectual property and cultural property interests. So, you know, let's let's work on that. Um, so partnership here refers to the establishment of genuine, respectful, and enduring relationships with indigenous people and entities that center indigenous leadership, priorities, and governance. It recognizes indigenous peoples not only as stakeholders but as co-creators, co-managers and co-deision makers in all aspects of the data life cycle. And that's why I mean when I say that formal contractual partnerships are so much more robust because it's too easy for people to slide out of these arrangements. Um because yeah, it's hard work. Um partnerships must be based in trust and shared authority and must reflect indigenous governance protocols and cultural values. Well, that's all well and fine until you run into the commercialization officer at your university. You see what I mean? Um and there are plenty of other people who want to raid our you know research outcomes and data. Um so this is a a serious matter for all researchers to consider. Governance logic partnership operationalizes recognition by embedding indigenous leadership and co-design into governance structures. It ensures that governance is enacted through indigenous agency not merely symbolic inclusion. Um partnership is the mechanism through which shared governance is realized in the humanities and social sciences and indigenous research data commons. It transforms institutional relationships and ensures that indigenous perspectives shape data practices. It's allergic to me. Doesn't like me. That's That's a cute little art cooperative out near the Western Australian border. um where we were working with some of our partners. Can you imagine what the archives of a an art cooperative looks like when you know they've been around for like 40 50 years. They have to have a record of every artwork. They number it, the provenence of the artwork. So like the indigenous data governance uh challenges for the ainal arts sector. Just have a think about that. Tremendous. But also the indigenous data assets held by little outfits like this right across Australia. Invaluable. So anyway, building capabilities. Now you know they have a new building. This is their old building. Um but they also have of course um their own database and and various platforms but I would at a at a I would hazard a guess that there might be two people in the community who know how to use it and if they're not around then nobody can sell a painting for instance. So capabilities refers to building the institutional, technical and cultural capacities required for indigenous communities to govern and manage data assets. This includes research infrastructure, you know, such as our cataloges for instance, um training in data management and analysis, legal literacy regarding data rights and research ethics and access to tools that support data governance. Governance logic capabilities act as enabling factors. They provide the foundation for recognition and partnerships to be effectively supported by resources and systems. Um and of course the conceptual institutional technical and cultural capabilities ensure that governance is actionable and sustainable in practice. And just it it sounds like word salad but you know it's a real thing. Just take for instance the language production center at the Uricolor school that we've been working on. So Kristen has created a catalog and everybody has to digitize materials and and catalog it. So then you know just take any item an elder has provided the story. um a a Yongu teacher has drawn the pictures and translated done the translations from Yongam to English, but they use four varieties of Yong and Martha. So which one um and then they always embed kinship and and um moyeti information. So it's either a ura story and therefore urae primer or it's a you know from the other moyete. So these cultural precepts are going to be relevant in relation to every item. And so that's why you've got to work with people because you can't just make stuff up. You put it into you put this item into a catalog. Without that information, you know, you you're losing information. the the item really becomes pretty meaningless. And we've all had that experience in archives, museums, libraries. You find something from say the the 1840s, no name of the Aboriginal person, no location of the place, but a a painting of an extraordinary person or object, and we have no further information. So we've had, you know, 238 years of that um kind of making us invisible, denying people's names, taking away place names, eradicating all cultural information. We have to reverse that whole process. So this is actually where in in building the capabilities of of our workforce that we pay attention to these matters. So I just wanted to give that example. Um and I'm sure you can think of lots more. Uh so self-determination as I said is the the effective right um that we should rely on here. Um the rights of indigenous people to access and use their data. It also encompasses the rights of indigenous peoples to define, control and benefit from data about themselves. This pillar includes the ability to set priorities, interpret data through indigenous worldviews, and refuse participation in systems that cause harm. Additionally, it involves the power to determine how indigenous data is collected, owned, and applied in research in a manner that aligns with cultural values and community aspirations. So uh you know there's been a a lot of conflict about uh data sets that are held by big data custodians. Um now the Australian Institute of Health and Welfare has advisory bodies. So does the ABS. They hold big data sets. When you're on an advisory body, it's really difficult. I'm not on one myself. Um, but I'm sure there are people in the room who have been on these advisory boards, but it's very difficult to do anything about that data other than to say, you know, there are problems with the data. We were talking about one of these the other day. Okay. Somebody wanted to survey Oh, yeah. Was it the ABS? survey our population again and put in a question about family violence and the Aberinal and Toouristra Islander Health Survey. Fortunately, Kristen picked this up and made the point that this is very dangerous for people. This puts people at high risk. Correct. >> Yeah. Yeah. So it's it's quite important in the context of closing the gap. I think it's um indicator 13 uh looking at family and domestic violence across for ainal and tourist straight islander people across Australia. It was a number of questions in the national aboriginal and tourist rate islander health survey. Um however when they were revising the survey for the most recent one uh they put together a an an advisory committee to look at that question and expanding on that question but they so that I I was on that that advisory group with a series of other aboriginal people and other experts um methodological experts and a variety of people but the one thing that we questioned was like okay so how are you asking these questions it's a household survey Think about it logically. If you're going to into a household where there is family violence being experienced and perpetrated, you're going to create significant harm for people in that household if you ask that question. So there are these ethical considerations that need to be dealt with in the way that we collect data in the first instance. So we pulled them up and said you obviously we need data on this this this really important issue, but you got to think about how you do this. Um so they did pull that question which is problematic because they haven't designed another survey which is another story. Um so that's just one example and you know we come across these all the time. Um and then of course finally inclusive the inclusive data ecosystem. So uh this refers to the way in which research data is created, stored, accessed, reused and repurposed within focus areas of the humanities and social sciences and indigenous research data commons. So this is where we have to accommodate uh cultural perspectives, different cultural perspectives, equitable legal frameworks, interoperable technical systems and collaborative research networks. um and prioritizing indigenous data governance needs while also supporting the broader house research community. So I'm taking up too much time on some matters but I did want to make this very relevant to you. So uh should I leave time for questions I'm monitoring? Well, you get it to work then. >> Sure. I'll try. >> Here we go. governance uh the governance guidelines. So, we've got a number of guidelines. Do we have time to go through these and the monitoring and accountability structure? Yeah, I think we'll just address this monitoring and accountability structure. All of the details in the the um the publication and we're going to have a workshop on it this afternoon. So we need uh deliberative, inclusive and and a culturally informed approach that centers indigenous leadership and governance at every stage rather than relying on abstract tools or metrics. Implementation must be driven by relationships responsibilities and so that involves institutional commitment and them saying it out loud. Uh so when organizations participate in humanities and social sciences and indigenous um uh focus areas, they must adopt the framework and embed its principles and data practices. Please don't think of it like a wrap. This is you know this is an operational document. This is more like a strategy. Uh in fact it is very much a strategic document to be adopted and committed to by institutions. Um so as I said indigenous leadership um because data custodians must be supported to lead the design, management and evaluation of data initiatives. This includes co-developing governance protocols. So this publication is not the end of it. you take this publication and embed it in your own institutions um and develop the uh local protocols um relevant to your data. Um and so this means moving beyond consultation towards genuine partnerships that share power and resources. Um and it involves capacity building. So you know we need to create a workforce to do all of this. It means that the way that we uh write our budgets will have to include uh budget lines on um indigenous data governance uh capability building um and all of the subsets of um requirements that will cost money. So and and in many ways, you know, this is better than this very disturbing government practice of randomly evaluating Aboriginal organizations and bringing in outside consultants. If you have data governance and your own monitoring and accountability structure, you can produce your own evaluations. So you know think about it as a living breathing strategy uh that you can now you know embed in your in your work and as I said you know the development of formal agreements that reflect indigenous governance such as data sharing agreements memorandum of understanding co-authorship arrangements protocols for attribution and providence and such agreements must be co-designed with indigenous communities and reviewed regularly to ensure they remain aligned with community values and priorities and the implementation must be adaptive. So the framework is not static as Jenny pointed out at the beginning and it must revol evolve in response to community feedback, legislative changes which we are hoping to bring about technological developments which are you know fast and furious. uh we're going to have to deal with AI. Uh but you know it it brings many benefits and let's have a talk about that later. Um so we need to be responsive to the diverse needs of indigenous communities as we implement this. So as I said focus areas must allocate funding, personnel and infrastructure to support indigenousled governance, training and evaluation. We must ensure that indigenous communities have access to the tools, technologies, and support they need to manage and govern their data in ways that reflect their laws, values, and aspirations. So many projects are not adequately resourced to support indigenous governance, training and evaluation. And this is the big flaw that must be addressed along with um the lack of legal recognition in the act and you know the subsequent problem that we don't have a single indigenous data uh custodian recognized under the act. So uh yeah indigenousled oversight evaluation and feedback and we really look forward to your feedback on this. So I'll stop there so there's some time for some questions. >> I was just going to quickly say um we will go through operationalizing this framework in the workshop this afternoon. I think it's at 1:15. So looking forward to sort of everyone getting their teeth into it. But just to leave it with um you know the framework itself is both a tool and a commitment. It's a tool for embedding indigenous governance in national research infrastructure and a commitment to uphold the rights, responsibilities, and aspirations of Aboriginal and tourist rate islander peoples in all data related endeavors. So, we'll leave you with that and go to questions unless Robert, you had a final >> question. See, you just look at them and they put their hand up. Uh my question is uh are there similar initiatives about other indigenous groups in the world? We uh well not me. I'm on the edges of it. I go to as many things as I can. Kristen and Andrea Clark have been attending um UN meetings online mostly. They went once to Geneva. There is a process in the UN but of course the membership is only member states that is nations and they're having debates about you know the very issues we're talking about here um and so around the world as I said indigenous people are excluded from the main game and uh we're trying to work with uh some people who are there trying to protect the rights of indigenous peoples you know in other countries like throughout Southeast Asia uh and so on and uh you know this is right now a hot topic in the UN. So yeah, we're probably better off than most. Um so to New Zealand, uh Canada and the United States, they're doing rather well. Um the, you know, the indigenous people. So it's in the settler states, the, you know, the English-sp speakaking settler states. uh where indigenous people are the most advanced with this. Um if if we can get recognition in the act in Australia and get an an enor set up will be doing well. Um but mostly indigenous people around the world are way out on the edges of this and struggling to cope. In fact, not just indigenous people, but African nations, I heard this morning from Kristen, are very worried about the impact of AI and they feel like they've been completely left out and they're thinking about what's strategy now, what's their strategy now in the face of AI because Africa is, you know, yet again uh a series of colonies from which the big players, you know, like Google are going to extract data. There's one data center in the whole of the African continent in South Africa. One data center. So that's the scale of the problem there. Hi. Um, is this working? All right. Yeah. Hello. My name's John Gray. I work for the South Australian Department of Health um as a data analyst. Uh my background is in social epidemiology and as you can imagine I'm often required to provide reports on things at very short notice and usually with very little opportunity to discuss them with anybody else except just do it. Um welcome to Adelaide. I hope you're enjoying your stay here. One of the things that I thought you might have some advice on is when we are trying to use large data sets such as the census where people have not in in practical terms given any permission to provide their data. They just divide because it's the law and Robert and I had a brief discussion about that yesterday. Do you have any advice about how to go about trying to get some greater sense of um authority for using indigenous data? >> A greater sense of >> authority for using that those sort of statistics. >> We're tackling that problem right now. So just to explain it to other people, health data uh in in order for somebody like what's your name? Sorry, John who's in the health department here to use that data to write the report that he's just been asked for. The data is is completely deidentified. So he's asking himself the ethical question, have any of the people who are behind this data ever given permission um for their data to be used in this way? Well, we're tackling that question right now in relation to the my health records. there's a push uh well a big project actually to um revise uh the conditions for the use of my health record data for exactly the reason that you're asking. So um we we're running three roundts. We had one in Cape York what a week ago two weeks ago we and they're virtual uh as well as in in person. So, if you want to be invited, write your name and your email address on a piece of paper and give it to us. We've got one on Monday. Is it Monday? Monday in Brisbane and Wednesday in Alice Springs and we can, you know, send you an email invitation to discuss exactly these issues. John, you'll have to come along to one of them. Um, so what are people saying to us about this? Uh well, first of all, in the ainal world, people are saying, um, well, I looked at my my health record and most of my medical information isn't there. It hasn't been uploaded. Um, and you know, of people in our our health sector, of course, say, well, yeah, of course, I I have no objection to my data being used, but it's really bad data. uh because the my health my my personal my health record doesn't have the time I went to get an MRI or this or that you know so uh and hang on I'll just hand over to Christian and some of the other things that we've been tackling in the context of the use of the my health record data is looking at whether consent should be through an opt-in or an opt out model. Um, and look, there's there's pretty strong push from the Department of Health to make it an opt out model. However, where our argument has been, okay, if if say you have very little digital literacy, you don't even actually know how to access or opt in or out of your own My Health record. Um potentially if you go to your local ato or your GP they might talk you through it but or you might have a family member who knows how to do it but potentially there's no one who will a explain what it means what the implications are and we're concerned that it really isn't free prior informed consent whether it's opting in or opting out. So these are some of the issues that I think are being debated nationally. As Marcia said, we're getting opinions from people around Australia at the moment on this as the initial consultation. There'll be broader consultation that goes on. Um but it'd be great to hear everyone's um opinions on these issues because it's a hot topic. It's not just going to be, you know, it's not just South Australia Department of Health. It's it's everywhere. So uh as I said we're we're trying to ascertain greater opinions of people across Australia in different contexts but I think there are a whole range of issues to do with data literacy. So I think we'll we'll be talking to Becky Cookie about about that as well. >> One there one here two more questions before little lunch >> I think. Yeah. um Julie Lee from Commonwealth Department of Social Services and that's that real ethical challenge about the difference between we have a significant responsibility to every Australian who doesn't have a choice to provide their data and often sensitive data and in a situation where they're providing it for a purpose that then doesn't lead onto the purpose that we then use it for. And whilst we stand behind we're doing it for better policy, better program development, where's the I suppose for me I get a bit passionate about we do lots of talking. We're responsible for the roll out of the Commonwealth framework for governance of indigenous data. But where's the where's the jumping off point? like where where do we start to actually do things and make a difference and see some tangible outcomes? Like we're really excited to see how your framework can work in conjunction with what we need to deliver under the Commonwealth framework and so excited about what we heard yesterday and what we're going to do this afternoon. but really challenged by the balance of the ethical the balance of so we're doing some work with the black town community data project and just trying to get the agreements in place so that we're appropriately representing community but delivering things with AI IHW on a platform that people can use there's no point in getting to the end of it and the challenge with target 13 and target 12 for closing the gap and for us target 4 and how we use target 4 to leverage to get the right information for target 12 and 13 and looking at things like how ABS do surveys how there's lots of pressure around longitudinal surveys like DSS runs the footprints in time and a few others but um how you ask those questions like as as a survivor like if someone had come into my house when I was still with my partner and sat there doing a household survey Okay, that puts me at immediate risk and getting people to understand what that actually looks like and how you get informed consent. Like I can sit here and say to you, Andrew, tell me all about your culture, your this, your cuz I'm going to put better policies on the table. What does that mean? >> Yeah. Look, fabulous question, a fabulous problem to have, right? But you know, we started out like asking these questions back in 2017 and here we are in 2026. Yeah. What nine years later finally with a framework for researchers. But we did we did contribute to the uh Commonwealth's framework. Right. So there just for everybody in the room to know there is a Commonwealth government framework for indigenous data governance. Um to be frank, we need to have more meetings like this with a mix of indigenous leaders, indigenous practitioners, our non-indigenous uh experts and practitioners, our epidemiologists, language experts and so on. We need to have more meetings like this where we can talk to each other freely about real problems. there is a a way in which these matters are discussed as if they're political and as if you know some slogans are going to solve the problem. We we need practical tools and I mentioned yesterday I'm not sure if you were here one of the um developments that I think really presents us all with a a great opportunity. Now, people are doing it all the time, but I'd never really noticed how useful it could be. And that is this, you know, urban indigenous health dashboards developed by a group of um indigenous community health organizations with professor James Ward at the post center in conjunction with AIHW. It's on the AIHW website, the urban indigenous health dashboards which brings together and analyzes and presents urban indigenous health data all completely deidentified of course for the first time. So see partnership indigenous governance. So the the health organizations provided the governance the expertise came from you know the post center and the AIHW. You get epidemiological um expertise and you get infrastructure because of the AIHW right so that's why partnerships work. So it's got that is a really nice case study that's got all the elements of this. It's indigenousled. self-determining. Um, it's ethical, it's respectful, you know, indigenous cultures there because of indigenousled governance. Um, and they're creating data ecosystems behind the dashboards. So, you know, there's a number of ways to go about it. So I urge you to stay involved with the ARDC meetings uh through our projects because likewise I think lang the language mob eldaka are coming up with some of these solutions as well. you can adapt solutions from other fields to your purpose. And maybe this afternoon when you're going through the framework and how it can p be practically implemented, you might want to, you know, get encourage people to come up with some other examples because there are some very good examples of people already implementing this because they figured it out for themselves. Sorry, that's the best I can do at the moment. One last one. >> Uh hello everyone, Hamid University lecturer here. Um I would like to know your opinions about generative AI software because I guess it is steals photos, paintings and other data of um uh indigenous Australians. Uh are you planning to push generative AI companies and not to record uh indigenous owned data or not? And if so, why? >> I think that the horse has bolted. You know, that's an Australian expression. Um, it's too late really. Uh, look, we should set out a a set of principles on that and people are doing that. Do do you really believe that um um Google's going to follow those guidelines? It doesn't work like that. There isn't some bloke sitting behind a computer saying, "Oops, there's some indigenous data. I better stop the, you know, the algorithm from scraping that." Now, you see, you see the problem? That's what I mean when I say the horse has bolted. Um what we need to do of course is make sure that indigenous data with our custodians has all the guard rails around it and that that that's the point where we put in the guard rails for the sharing of indigenous data because the you know Google and all of these companies are going to scrape data not from our little organizations. They never put their data online anyway. They're going to scrape it from ABS and I think ABS has already been cyber attacked at least once, right? Yep. So, uh, now ABS data is deidentified, but you do panic when you hear about a cyber attack on ABS and think, I do hope they've got that right. So, you know, this is where we have to be vigilant at that level in our partnerships. Let's make formal partnerships with our data custodians and that includes the police and health departments. We need formal partnerships to make this work to protect ourselves from AI. But on the other hand um you know we need another meeting sometime to talk about the benefits of AI. There are some people who are saying don't publish any data. Um let's pull our journal articles. Uh because there's a risk that when you publish your data in journals that you know the the baddies are going to come and scrape that data and use it in ways that it was never intended to be used. Ah I can see an AI expert tentatively putting his hand up. Way you go Nick. >> Thanks Marcia. uh Nicholas car um working with the IDN just a point on the what to do about um generative AI generating indigenous art and all that kind of stuff. So one thing you can do is to make sure that your own images and videos and text and so on have good metadata. This sounds lame but it's not lame. Um when if you remember when chat GPT in these systems first came out and you did searches you would see results and then very quickly people are like oh my god where are these results coming from? So when you look at those systems now they actually do have source attribution to certain extent. Is it perfect? No. Is it better than it was a few months ago? Yeah. So if your source information that is being used by anything, I mean generative AI is one thing, but if your source information has good metadata, uh there's a higher likelihood that um the uh information about origins and so on will flow through. So imagine we saw some uh AI generated indigenous images or images of indigenous people yesterday. Um, imagine in the future that there's some significant um image of indigenous stuff including synthesized art and so on. Um, it wouldn't be unreasonable to think that um that there would be protocols in place or interest in the digital provenence of that and what you'd want to see is look, I've got this picture of an indigenous synthesized picture of indigenous people. what was the systems source to make those imagery uh that image and um the better we can um annotate and catalog the original source material that these things use the better the annotations are well the more able the annotations are to flow through to the endpoint we'll see how we go with regulation and so on but >> so a question Nick do our metadata tools on on our indigenous data commons help you to do that please answer the question >> uh for digital objects as a whole. Yes. For individual images, no. We could say, and I think that this is probably sensible, we could say that we need a tool to help annotate images in particular uh with content because sometimes images embed metadata. So, what we could have is a little tool that says, hey, have you got an image of an, you know, indigenous artwork or something? Do you want to embed like a digital image? Do you want to embed metadata in that image so anybody who uses it might be able to see that metadata? We could we could implement a tool to do that. >> Great. Thank you. That's what I want to know. >> I was just going to say that this is definitely something that's being discussed and debated internationally. I was also at the United Nations data working group on data governance meeting last night where a lot of the member states were discussing this very issue. And there is actually I think coming up on the 6th and 7th of July there's a United Nations um global dialogue on AI governance. If anyone's interested in going along we'll be paying attention. I was just oh it's Cat Cqai from Charles D University. I was just going to add that um we already do have guard rails that uh you can't enter the prompt I want a Marvel image of myself because Marvel commercially has blocked that. So there must be a way we can block a request for an indigenous version of whatever.