IGF 2022 Day 0 Event #48 Data Stewardship, an alternative governance framework for Data Justice

The following are the outputs of the captioning taken during an IGF intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.



>> MODERATOR:  Thank you for joining us, everyone.  We just need a moment to make sure everyone can join.  He is joining remotely so we're just making sure he's able to be on.

>> MODERATOR:  Hi, everyone.  As we wait, maybe since we have a captive audience it would be great to know who is in the room, and if you would be open to introducing yourself or why you're interested in the session.  We would be happy to hear a little about your backgrounds as we get started.  We would love to hear from who is in the room and if you would love to volunteer an introduction, that would be great.


No pressure.

>> (Speaking off mic).

>> MODERATOR:  Fantastic.  Thank you so much.  Anyone else?

>> (Speaking off mic).

>> MODERATOR:  Fantastic.  We've had a conversation with policy as well.  Good to see you in the room.  Thank you for coming.  Anyone else?

>> Hi, my name is Nia.  Senior research office ‑‑ (Speaking off mic).

>> MODERATOR:  Fantastic.  Thank you for being here.

>> Hi.

>> (Speaking off mic).

>> MODERATOR:  Fantastic.  Really important work.  Hi.

>> (Speaking off mic).  Data governance ‑‑ that's why I joined the session.

>> MODERATOR:  Thank you for joining.  Great to hear your insights as well.  We're just doing short introductions as we pull up our presentation.  We're having some technical issues, but we would love to know who else is in the room.  Hi.

>> (Speaking off mic).

>> MODERATOR:  Thank you so much.  Any last introductions if you would like to introduce yourself?  No pressure.

>> (Speaking off mic).

>> Hi.  My name is Tijani B, Chair of the MAG of North African IGF.

>> MODERATOR:  All right.  We're going to get started in just a minute.  Our apologies, we had a colleague joining remotely, but it sounds like he's unable to be on video today with us.

>> I'm sorry about that.  Just to repeat.  I'm sorry, we had a colleague that was meant to be joining us on Zoom but is unable to make it or is having technical difficulties, so that's the reason for the delay.  We are really thrilled to have you in the room with us.  Thank you for taking the time.  I know it's been a long day.  We appreciate you stopping by for this evening session, and we're going to get started in just about a minute or so as soon as we have the slides up.

I also wanted to mention that we are keen to sort of keep this dialogue open and would love to hear your perspectives and expertise.  I think you come from various backgrounds, so we would love to see how we can collaborate with you in the future, or see how we can mesh these insights from today also in the work that you carry out in your respective jurisdictions.

Thank you for being here, and we hopefully will have an interesting discussion.  I also want to flag that we want to keep this as interactive as possible, so we will be presenting case studies as well as insights, but at any point, if you would like any clarification, we would love to hear from you and want to keep this as open‑ended as possible.

>> Is it okay if I ask you to change to the next slide.

Thank you again for being here, everyone.  We're going to get started.  I'm really grateful to be joined by my two colleagues today.  Amrita and Ketosh from ICT Africa and we don't have Roland on call, but he's also participated in putting together this material.  So, I will start out with a quick overview of what we'll discuss today.  If you can go to the next slide, please.

So very broadly we'll do a quick introduction to our organizations, talk about the challenges we're seeing in the way that data is collected and governed in the data economy today, touch on a few case studies where we're seeing innovation in the way data is collected, governed, what the implications are for the policy environment, as well as where participation fits in.  And then we do have a section for resources that if we do have time, we'll get into.  Otherwise, we'll share this deck with you, and hopefully be able to take forward the conversation on that front, and hopefully we'll have some time for questions and discussion toward the end.

Wonderful.  I'm going to hand it over to my colleague to introduce Aapti.

>> AMRITA NANDA:  Hi.  Everyone.  Mime Amrita.  Hopefully you know of Aapti Institute.  I see a few of you in the room.  We're based in India and work on everything at the intersection of technology and society with a focus on how people actually negotiate access, inclusion, and sort of look at this through two lenses, both the online and offline.  So, the team that I work primarily on the Data Economy Lab looks more at ‑‑ looks more at policymaking, technology development, the ways in which community is actually negotiating with digital interventions or how they're negotiating with increasingly data in the world.

The second article, the Digital Public Lab looks at offline questions of access, so what does last‑mile access look like for communities, and what does offline infrastructure need for everyone to sort of be included in an increasingly data‑full world.  I'll hand it over to Keto to introduce Research ICT Africa as well a little bit.

>> Hello, everyone.  Thank you for joining us.  I know it's the last session and everyone is having a brain but we appreciate it.  I'm Kristophina Shilongo, and we conduct research with facilitating evidence base and informed policy.  Our research is also public interest and mainly on the digital economy and also on how society responds to it.  We aim to contribute to digital equality and data justice, which is a very important part of this presentation and decision today, and for the session today, we are presenting our work under the Africa Just AI Project and we have two main themes we're looking at, beneficial AI and sustainable development, as well as coming up with collective and participative governance frameworks with the intention of contributing to data justice.

We will explain what data justice is throughout this presentation.  Thank you.

>> AMRITA NANDA:  Thank you so much, Kito.  I think we've all collectively come to IGF with the understanding that there are a few things that we need to solve in the room, and I think a few of the challenges we highlight today as being lack of transparency and accountability, so a lot of data is collected, but we don't have an idea of what that looks like as individuals let alone as communities.  So, it's extremely hard for us to be able to know what data is collected and what value may come were the data.

The second challenge is that we don't really participate in questions around data collection at that individual or community level, so it leaves us with little option of how to collaborate or how to connect with other people, build solidarities around data, learn how to generate collective or shared value because often the data of value is often relational and often just sort of siloed in the Internet or digital processes at the moment, tend to silo us as individuals in the ecosystem.

The last issue is that often a lot of this data is held and controlled by certain entities, often larger corporations, could be even state entities, where they're sort of unsure of how to release this information or don't have the incentive to release this information.  And in the case that they do, it becomes extremely hard for others to use it, so you run into other challenges with respect to how to generate value from the data at an open level.

So, we see a few opportunities that arise from the challenges.  One that there are opportunities for individuals to play a bigger part in the way that data is collected as well as governed, and we see this happening through stewardship which we'll talk about very soon.  The other is thinking about how we can collaborate with multiple entities, whether it's state led, the private sector, Civil Society, important to have in the room and build data collaborations with.

The last is also how we can think about these levels of technology, both at an online and as well as offline level, so as we're transitioning to a digitized world, how do we also have processes that are offline that are running in parallel to ensure that things like access are not sort of assumed for those who don't have access to sort of digital infrastructures and technologies.

So it is on this basis, if you can go to the next slide, it's on this basis that we believe that a new just and fair contract for data is required, and we see data stewardship which is an umbrella term you'll hear us unpack very soon, as being at the heart of this.  We essentially believe that stewardship can help facilitate a lot of these larger themes around user agency and participation, around transparency in terms of safeguarding digital rights when it comes to privacy, cybersecurity, et cetera, as well as unlocking instrumental value.

These are big promises, but we'll jump into how these are rooted in the framework of data justice and my colleague will take that.

>> KRISTOPHINA SHILONGO:  I think a lot of discussions about data is that we want to unlock the value of data, specifically in Africa, for instance, there is you know a lot of people coming up with solutions or like frameworks on how people can use data to provide solutions but as economic solutions and data justice is one of those paradigms which facilitates it.

And, of course, we want to say ‑‑ we want to take a step back and say that, hey, we may have data but then this data is also rooted in discrimination, structural and systematic inequalities as well as also keeping in mind that it furthers discrimination online and offline.

And for this concept, experts have developed three kind of principles that deal with invisibility, and invisibility of data subjects or data communities and also looks at how communities engage with technology, so when you build this technology just based on these data, how do the communities engage and also considering how they disengage communities as well.

It also looks at discrimination and how to facilitate and make sure that it is not ‑‑ it is anti‑discriminatory.  And, of course, a leading expert in data justice, envisions data justice promotes fairness in the way people are made visible and treated as a result of their production of digital data, and this is necessary to determine ethical pathways through a data world.

And, yeah, so this is seen ‑‑ we see data stewardship as a means to data justice.  I'll hand it over to Amrita who will talk about data justice.

>> AMRITA NANDA:  Just to give a primer as to how we see stewardship as a framework.  We play part of an introductory video that we have.  These are resources that you can find on the website or ask us to send over.  Can we go ahead and play the link?

>> (Speaking off mic).

>> The video aims to visualize where data stewards will be positioned in the ecosystem of data sharing.  Are there any questions that you might have of what we mean by any of these terms or sort of how we're envisioning the stewardship?  Just a bit of a check‑in.

>> Maybe we can just talk through it.

>> (Speaking off mic).

>> Okay.  Give it a minute.

>> I didn't realize.  Maybe some of the attendees online through the Zoom links have questions or have something for the chat?

>> We'll make sure to share the video later.  Essentially, given a lot of inequities that have been outlined and broader vision of increasing data justice and more for fair and equitable data economy, we envision data stewards as trusted intermediaries in between users, fiduciaries, and data requesters.  Data stewards can look in many different ways and we'll talk about the frameworks we've seen, but essentially, they can play a host of roles.  The roles we see here are not exhaustive, they're also not always combined within the same steward.  We could sort of picture them adds both technical and structural.  From a technical lens, the stewards can play the role of increasing data literacy, and even building standards and frameworks for data use, playing the role of an analytical layer, allowing communities to understand the data better ‑‑ excuse me.

From the sort of structural lens and this is a very important one and one we'll focus on in the presentation especially is the way that data stewards can facilitate participation, so increasing the agency and autonomy of individuals in negotiating with how the data is used, how it's processed, or even having data visibility over it.

Finally, they can also offer a grievance and mechanisms we see now exist within privacy laws and frameworks, but may not always be actionized on the ground in the way we would want them to.  Next slide.

As we do research over the past couple of years, this is a lot of primary research by speaking of primary stewards or but as secondary research and sort of landed on what the ecosystem perhaps needs to look at to support the efforts, to support in general more human‑centric data governance.

What we found is that while at the top layer are domestic international regulations and policy measures that are typically state efforts, that needs to be married with the sort of bottom‑up understanding that ecosystem enablers, which I think is a path you can't see over the captions, but being the technologists, builders, the people actually working with technology on a daily basis, and the sort of second layer of this is the philanthropies, the people who are able to fund research and fund initiatives like this where communities are unable to.

So, what you see here is sort of a consideration of the taxonomy that we have so far of the kinds of models that we've seen.  They all differ in the design choices, and the ways and degrees to which they encourage participation.  And the kind of benefits that they are geared to.  So, the way that we see it there is no one‑size‑fits‑all for either of these models, but different data types for different sectors and definitely for different jurisdictions, different models work in combinations.

So, we've been sort of focusing on climate and environmental data as a sector, so I will talk a bit more about that.

>> Okay.  So, you've heard from Amrita we've seen the data constellation, cooperative, exchanges, and many of the terms are familiar to the people in the room, I imagine.  We're seeing how this aligns or applies to the collection of environmental data or data that has to do with sustainable development.  It's something that we've all had to collectivize around and it's important for us to determine what the different model types are and what the approach it's may be as we think of this for policymaking efforts and for kind of reconciling data from different players and from different sources.

So, we'll go on and we'll talk about a few different approaches, but before that, I want to talk about participation a little bit more.  You've heard us say a participation a few times, and we see that this can vary across thinking about how this can be designed, so as we see on screen, there is a few different layers or levels of participation, right.  So, depending on who the audience is, who the beneficiary community may be, you may be engaging with them at various different ports, informing them of how data is being governed to being able to empower them with actually taking decisions on who data is shared with and for what purpose.  This can range, and we also cover a few different case studies that you see the diversity in participation.  It's important to consider this slide of the sort of ladder of participation because as you're designing, whether it's stewardship or data‑collection efforts, you will have different entities, whether they're beneficiaries or organizations that can engage or participate at various points, right, and may have access constraints, may have constraints with respect to their own data and digital literacy, as well as their incentive to participate.  Not all of us actually want to engage with questions around our data.  Maybe we want someone to advocate on our behalf.  As we keep this in mind, we want to go through a few different case studies.

>> Awesome.  The first is a high‑level of participating.  We see this in the context of agriculture.  This is a project that we're actually running along with two other organizations, so the 17 Rooms Project funded by Brookings Institution as well as with a cooperative, so an agricultural cooperative based in India referred to as Mecca and this follows under the broader association or Sava and central point of the question or point of research has been is what is sort of value of generating a data layer on top of an existing cooperative?  So many of us are familiar with cooperative structures, how they exist in sort of offline settings, and we were asking is there value in actually placing a data layer on top and helping pool data that the cooperative generates and then being able to deliver these insights back to the women who are part of this cooperative?

So, we had a few different questions that we asked.  One was around the actual governance of the data, so how would women want to be involved in this, are there existing structures around decision‑making that exist within the cooperative that can be taken advantage of, what do those look like?

The second question is incentive to participate?  Is there value that women see in being able to pool the data together?  Do they see any benefits together or how can they be shown the value if it exists?  Third is the data itself, what data can be collected from women, the activities they generate, or from the corporative itself in the running of the cooperative.  With respect to that, what is the technology needed to power the decisions or governance of the data itself?

Lastly is the capacity, we mentioned data and digital literacy a lot.  When you think about the specific group of people, what is the engagement that, for example, women farmers need from a sort of capacity‑building standpoint to better understand how data may be able to be beneficial for them, and how do they relate to data in their own terms?

And so we've seen that there are a few learnings around this, so when we interviewed women and our colleagues went out into the field, there is a few ways in which data or pooled data can actually be delivered back to these women.  The first is that it actually enhances their creditworthiness because there is a foundation or layer of data that becomes available to share with financial institutions.  The second is that once they have this information, or insights from the cooperative itself, they may be able to scale the actual activity of the cooperative and generate greater profits.

The third is there is greater sort of autonomy and there is visibility into the sort of ongoing of the cooperative itself and are better able to take decisions on that basis as well.  So that is the cooperative structure which is a high level of participation, as Amrita described, these exist offline so this is really looking at how this can be translated into an online setting, and there are some great opportunities we're seeing in this space.

The second model, if you can go to the next slide is a data collaborative and speaks to the point of meeting stakeholder to come together.  Data collaboratives are often organized on basis of shared mission or purpose.  In this sense, this data collaborative brought together stakeholders from the public sector as well as from the private sector, we see refer institutions being involved and multilateral organization and that's the UNDP.  What they ended up create something a digital public good.  As you see on the screen, there are a few screenshots of what that looks like, and it's actually called the data for climate resilient agriculture platform, and we're happy to share sort of resources on this as well, but it's interesting because it compiles existing open data, so geospatial data and they also worked with the governments to be able to pool datasets that they possess and actually visualize them and make them accessible.  There are various ways and toggles to customize to see different views at a regional level and also at a global level.

What's interesting about this kind of platform is that you can see insights at a micro level to really toggle the filters and indicators or indices and look at broader historical trends and patterns which is important for foresight or anticipatory policymaking as we go forward.

The other important element of creating these digital public goods is they can be customizable.  In this instance, you're seeing this represented in a state in India, but if you were to partner and create a data collaborative with a government in Cambodia, then you're able to pool data that exists from an Open Source as well as public source and get some really interesting insights that come out of this.

And going forward there is a few other models that my colleagues will also speak to.

>> Can you go to the next slide.  Thanks.  To this next case study, I think is really testament to how different types of stewardship models or intermediaries can come together to create the larger sort of purpose.  Two examples here.  One is SAFETIPIN a technology platform launched in India that functions in a number of cities across the world in Kenya and Colombia.  Essentially, what the app enables users to do is to audit public spaces themselves, targeted specifically at women and girls, to be able to develop safety scores for the public spaces, and what's especially interesting and I think we missed the screenshot is the kind of parameters used on the app, actually speak to a lot of ‑‑ a lot of things that we as women contend with when deciding whether a public space is safe for us or not, but aren't often quantifiable, aren't something we're able to aggregate and think about, like lighting, openness, visibility, how many women we see in the area, gender division, and there is also a sort of second part to this app which is called SAFETIPIN Light that collects visual data every 30 meters using an IoT camera that is attached to a vehicle.  For this there has also been partnerships with private and public sector within SAFETIPIN.  And why this is interesting is when combined with the Indian open exchange, which is an Open Source aggregator for Open Source data and mobility data, data from a source like SAFETIPIN is well contextualized by automobile data and also the IODX serves to come to sort of build standards and ensure that data input from many sources, not just something like a SAFETIPIN is standardized, readable, and in turn usable by third‑party innovators, so whether there is startups looking to build citizen services, but as for urban governance decisions, so the data is then exchangeable between government departments, different state departments, and so for the users themselves, there is a much more ‑‑ there is more diverse and sort of evidenced knowledge on how to navigate public spaces, and this is particularly true for women.  It sort of creates this aggregated standardized set of insights for policymakers to make better decisions as well.

This is one of the ways that we've seen two sort of different models come together, and a number of these examples exist throughout the data exchange as well.

We'll talk a bit anyway on data cooperatives.

>> KRISTOPHINA SHILONGO:  Okay.  I'm going to speak to some examples from the African content and specifically Southern Africa and one of them is the Abalobi, social enterprise digital platform that was founded in 2015 that connects fishers and buyers, people who are fishermen and as well people who buy fish.  It presents a sort of publication that helps fishers to manage publications and sell catches, and beyond that also used as an advocacy tool to also ‑‑ to advocate for different policy interventions.

So, this platform is backed by a participatory action framework, and it aims to protect the rights of small‑scale fishers in Africa.  If anyone is from Southern Africa here, the fishing industry works in the way ‑‑ the fishing quotas, for instance, and most of the time the big enterprises are the ones who have more rights at the determinant of the scale‑scale fishers, and so this platform or these applications, we see that the stewardship model has three users with the data contributors, which are the fishers and everyone else who works with them.  Abalobi is data collectors and then research facilities and as well as government and then specifically the Department of Agriculture and forestry and fishery.

What we've seen with the project is that the kind of impact very much aligns with the justice tenets and principles and one of them is based on the data collected, the community, the fishermen are able to have collective negotiation and bargaining.  So they, you know, they can use insights from the data that they collect to talk to, you know, different departments in the government.  They can also advocate for their rights.  If they're fishing in a certain place and some sort of corporation like a mining company comes and wants to extract or place an oil rig within the region that they fish and they have the data to say that hey, we have this place, it's beneficial to us and we have the data that it's providing economic opportunities and benefits to us.

It also, yeah, we've also seen that other stakeholders can come and request for data, which also fuels into innovation, so people can ‑‑ third‑party requesters can ask for insights based on the consent from the community, it is given to them to feed into sort of like innovation that they might have.

And secondly, it also makes visible labor that is traditionally invisible.  They have seen for instance that women assist a lot of fishing activity that happens show and you see the fishermen in the see and then the women that are packing the fish, the work is very much invisible, so they have a points‑based system where, you know, when someone, for instance, packs a box of fish or sells it to a local shop those activities are recorded and counted up and they have some sort of a point system for it.  They get data ‑‑ yeah, the data is collected and renumerate it because of the fact that they also get insights to trace how the supply chain works.

They also empower communities with data, so these insights can help, you know, the fishermen to see, you know, sustainable ways to catch fish, for instance, or sustainable ways to catch fish, and also improve on the supply chain as a whole, and it's also safeguarding digital rights.  You know, it's beyond the traditional way of just ticking a box and saying that we consent to it.  They provide visual consent mechanisms, so they provide ways of, you know, like infographics and saying this is hard work.  They also have a very strong participatory decision‑making framework.  They have a board, a nonfor profit board that governs the way the data on the platform is shared.  If anyone has questions at this point?

>> We've gone through a lot of case studies in different sectors and different models, so just wondering if anyone has questions on any of what we've sort of talked through so far?

>> I also want to preface we talked about different use cases and sources of data will differ as well as use cases involved.  With data collaborative for instance, you have various stakeholders to come together to create a various technical infrastructure.  We see for data and climate agriculture, but you see the more small‑scale applications can extend down to helping farmers to realize the value of their data or fisher folk realize the value of their data.  The use cases for the stewardship can be on the size or extent of the problem looking to solve for or type of data that needs to be collected.  I want to preface because there are various ways data stewardship can mean.  We do have a legacy of understanding data stewardship very differently within organizations, so the framework we're presenting today really deals with things like the governance of data, as well as the more technical facets of how it can be safeguarded, so thinking through things like encryption, thinking through thing like other process it's like data minimization that we see from policy standpoint, so encompasses a broader realm of discussion around data governance, not just technical safeguards but as data at a human level in some ways, which is why the case studies that we've chosen kind of highlight the more individual or community‑oriented possibilities in some ways.

So, I wanted to pause here and, you know, as Kito also described, this would be really interesting to get your opinion on where you see this land in your respective jurisdictions, in your work, data for policy is a huge buzz word today, so we would be keen to see how this is applied.  Any questions?  I'm happy to hand over the mic.  I know we have another one as well.  We'll use this to pause here.  Anyone online that would like to jump in as well?

>> (Speaking off mic).

>> I'm sorry.  I can't hear you.

>> Yes.  I see your presentation, but my question is, do you apply any pilot project into the election, or maybe political campaign about the data established?

>> So, we haven't personally looked at any data‑led around campaign.  I know there is quite a bit around identity and access management and I think that would feed into the sort of campaign process, but it would be like one of the building blocks.  I think we're seeing in the broader conversation around digital infrastructure and public goods how public sector as well as maybe private sector can create the products that link identity, create access, and also create platforms for citizens to be able to interact with the state.  So, we're seeing stewards sort of fit into the broader, you know, puzzle piece, but not with respect to elections specifically.  Happy to take it offline as well and see if there is, you know, new examples that we can pull from that might be able to address some of these concerns.  Any other questions or?  Hi.

>> AUDIENCE MEMBER:  Thank you.  Talking about what you just mentioned about the different users of the word advocacy ‑‑ I'm sorry, I was going to talk about that.  Stewardship.  I think the work at that we've done, we've tried to talk about data stewardship, but to drive the conversation away from data governance but at a country level and not as much as an advocacy role of individuals going to individual scales and use of data in an individual, which I think is what the case studies represent, and then I'm wondering if you're using the word wrong, because what we're trying to do is to help governments move away from a concept much data governance because I feel that's very related to data sovereignty and that comes along with all of this expansion of data localization measures that are impeding the sharing of data across borders, so we're trying to change the language from data governance to data stewardship to help governments understand that the important thing is to use data to extract value for social and economic value of data, but as share it with others.  But are we using the data stewardship word wrong then?

>> I'm happy to start and if my colleagues would like to add on.  That's a fascinating question, and I think a lot of cases that we've chosen today, as you mentioned, really focus on the individual or community value question and we haven't looked at the sort of possible uses cross border in some sense, but I think the way that we think about stewardship is kind of overarching.  We also are aiming to kind of push the envelope on data governance a little bit because it ends up being very organizational specific or very state specific.  And so, yes, policies are important to consider as well as data privacy and data governance, but I think there are use cases for stewardship at an ecosystem level, which is what we're trying to push for, and the reason we chose sustainable development and environmental action is that I think it is cross‑border in some sense.  So, the question around digital ‑‑ I'm sorry, data for a climate‑resilient agriculture is one that will have to be addressed by neighboring countries, by our global sort of collective in some sense as well.  I don't think we're defining it very differently.  I do think it is maybe a result of some of the cases that we've proposed.

I think there is also enormous value in looking at sort of global issues that we have to be able to contend with together whether on issues of water management, issues of air pollution, all of the indicators far more valuable to have this data shared amongst actors and countries as well.  So, I don't think we're advocating for data localization in that aspect either, but I do think we're envisioning it at a very similar level and I will also leave my colleagues to jump in here.

>> I agree with you on the question of the definition, but I think through some of the research that we've done on what data localization can look like and whether that's really at odds with community value, for us one of the important questions was what does stewardship look like for nonpersonal data versus personal data and what are the kinds of taxonomy that different countries are going through or arriving at to be able to make cross‑border data flows a reality.  So I think that's where like the tension with data sovereignty comes in, but I think it maybe is also helpful, for example, we wouldn't see a cooperative sort of ‑‑ or at this stage engaging in cross‑border flow, but at the level of thinking about open data mechanisms or what nationwide data exchanges can look like, that's where the sort of mindset of orientation of stewardship is something that is oriented towards public interest, public value, and finding ways to define the public value across different data types, that's maybe one way of transposing how we think about data stewardship to this sort of more international thinking.

>> I just want to add that I think in Africa we also have the same kind of thinking, and the response of like trying to break down, you know, data governance through like data stewards and these different like tenets is because of that, you know, speaking of data governance, we've always like I think people initially were talking about it from a national and also very much pushing for data localization and data sovereignty and like there is no value in like data that is stored in your country.  Like the value comes in how it is used and also now when we introduced the concept and that's why we want to really like push this framework of justice to say that, yes, data ‑‑ you can have data, but like as well like when you use it, it can further injustices for other people.  So, we're trying to like ‑‑ I don't see harm in using data stewardship in that method, and I think it points to the fact that policy frameworks on a national level and continental level should respond to what's on the ground.  So, you know, data governance policy frameworks, should help facilitate this collective or participatory ways of handling and processing data.

>> I think you had a question?

>> AUDIENCE MEMBER:  Thank you very much.  The session is very interesting.  I've learned a lot.  Thank you so much once again.  Now we are talking about data, but I think I'm a little bit confused because to talk about the data, how many of our countries or population are value or knows about the data?  What does data mean?

You see, even if you come to Africa or in our country, if you tell them or talk about data, they don't care because they have no value in what data means.  First, I think education should be there and I think after that also the government should care for citizens, for country, about the data, so this is the first main problem.

Then after that, I think we can talk about data storage or other things, otherwise, without educating the population, our government, in a sense let us take different NGOs, NHOs, they have different software, they use a Cloud system, most of them are in the states, so they get employees better, but your government doesn't take care.  They don't care.  So how are we talking about data stewardship without, you know, treating awareness to our government or population.  Thank you very much.

>> I can start us off.  No.  I mean we definitely agree with the first premise, the first layer is digital literacy, that kind of advocacy, and sort of even just diagnosing what can incentivize communities to think about data in this way, to think about it ‑‑ to think about the data rights in a more holistic manner.  And whether also if communities in certain jurisdictions have the bandwidth to take on say a consent burden, or even the burden of sort of participating in a cooperative and doing that sort of decision‑making.  So I mean I will preface this by saying that this is always an open question for us as well, and it's something that in the way that we envision stewards and some of the examples that we've seen as well with like making an existing data cooperative, so a situation where the communities who already have a governance framework where decision‑making is allocated in a certain way, but how can we add a data layer to that and do exactly what you're talking about to sort of both add capacity, build literacy, and I think also with the ABALOBI example, the way in which consent notices are given now, so rather than your traditional not very effective, or not all effective reams of text that you check a box.  So, sort of thinking data literacy itself, does it have to be a sort of mammoth of capacity building or can it be communicating with communities in a way that just requires less of that upskilling and capacity building.  Yeah.

>> I think like Amrita said, that's a very important question.  And I think for us, many times when you talk about data, we talk or think about, you know, that data is collected from all of us and therefore we think that we should be kind of like on us that everyone should know, that we use computers, and we don't know how the motherboard or CPU works.  Sometimes I don't, and maybe the concept of data stewardship questions like should communities know everything, should people who are subject to injustice know how these models work?  Should they know the nitty‑gritty of every governance model?  Do they have to know?  Is that a burden that should be placed on them?  And one of the, he thinks, these examples, for instance, they identify.  We always talk about in the circle, we talk about stakeholder and stakeholder responsibilities, so this is a new concept, you know, and we're all trying to learn.  So, we're also identifying the different kinds of stakeholders and their responsibilities, and Aapti has come up with a very interesting playbook that identifies all of these different stakeholders and roles and their responsibilities.  I don't think it should be, yes, data literacy is important but toward who?  Who is the end user?  Who are the programs aimed at?  And on the value, and I like that you said the value of that, and I think you don't sometimes, for instance, there was another case study on Blue Forest Project in Mozambique, and the value is not in how it's happening, how the data is used, but the value is perhaps also like when someone comes into a community and provides like this technology and technology beneficial, it's gaining the trust from the communities and saying that, okay, you have this data that comes from satellites or whatever, and we have this knowledge that we've been living with these like forests or these trees for centuries, and we're collaborating.  So, this technology is designed with our needs in mind, so it's not per se that they even have to understand data, but it's just understanding how that came about.  I think stewardship contextualizes and identifies and understands who is involved and what is the outcome.  I think that's the most important at least in the work that I've seen, just allowing like trust within communities and trust between different stakeholders.

>> I think that was really well put.  I do have an additional point to add but I want to open the room to see if there are reflections or reactions to that.  I know we've been explaining a lot, so if there are any questions or thoughts on this, we would love to open it up?

>> AUDIENCE MEMBER:  Thank you very much.  Rod knee Taylor, Caribbean telecommunication union.  I joined the session late so forgive me if you maybe answered the question before.  When I arrived, you were talking about data localization, and I want to link that with the data sovereignty as well in particular because we advise governments on policy, right, ICT policy in the Caribbean.  And for us as in the Caribbean being small I say developing state, much of the infrastructure Cloud services are built in other countries, Azure, AWS, Europe, and for us we're pushing, of course, things like civil registry information, birth certificates, medical records, criminal records, and so from us, from our perspective, there is an issue in having that kind of data outside of my jurisdiction, not just from a national security point of view, but certainly most of our islands are connected to these clouds by fiber occasionals, and if there is a disruption does it mean the government service is shut down and I'm no longer able to access my teleregistry, so we understand the value of having data externally from a resiliency point of view and so on, but what are your ‑‑ what have your studies shown, or is this a real concern for us, the small island developing states, or is this, and how do you address that?  Let me put it that way.

>> So, I think I would be best situated to answer that at a principal level and not so much diagnostic and what you should do.  I think whether it's a real concern or well‑founded concern or not is relevant because it's something that all countries are thinking about, so we have to sort of contend with it.  So I think the larger question, or sort of this is what our approach is, and at least in non‑personal has been is to look at localizations on standardizations and how you're even thinking to the future of a country's data governance is to think about the ecosystem as a whole and see sort of per jurisdiction what does it take to create data regulation that can be layered, that can evolve over time, that can not necessarily pivot from one end of the spectrum to the other on matters like data sovereignty, but being able to dynamically evolve with what we learn going forward.  So that's ‑‑ that's also work that I can share with you because we've done a sort of scan of ‑‑ and this was last year, but of countries and how they're going about at a principal level what the data sharing looks like to different countries, and based on our analysis of where most of the benefits lie, we saw either and even at the level of the country itself, neither the mandatory nor voluntary data‑sharing mechanism, but one that enables the ecosystem as a whole.  So, incentivizing and sort of injecting these enablers through regulation into the ecosystem.  So that also, I think the key is that it allows for room to develop and room to make changes, for regulation to be fluid, which is not something that we've required not past, but with a field like data and something that changes so dynamically, we're seeing that that should be sort of the orientation for most policymakers.

>> I was wondering if I could add something.  Because the data legalization is something that I worked a lot on.  I'm Natalia from CBI institute and as part of the government we've been analyzing the benefit and hinderances that data localization measures can cause and what do we have perhaps seen is that it's important for countries to try to do data specification, so there are certain data that is too sensitive to share or to even store abroad, especially with the U.S. Cloud Act or the power ‑‑ well you have either U.S. technologies and Cloud bases or China, which also has problems of government use of data.

So, with that data classification then, we have been able to try and show countries how they can classify their data, and depending on that data classification, then take decisions of where to store it, either public Cloud or physical data centers that are not linked to the Cloud.  But as something that we have found very interesting is, for instance, most countries consider health data very, very sensitive, so they restrict processing and storing of data to their own borders.  However, we have found that this hampers health research enormously, and health data that is anonymized doesn't work.  It's useless for research.  So there needs to be room for exceptions as it was, for instance, with what happened with COVID‑19 that there was an exception to these extreme protections over health data to benefit research and to benefit like public safety and public health.

And if I may just one final thing, what we have also found is the importance of governments and trying to understand how they can collaborate within certain priority areas that they have within their own region or with other countries so that they can work on sharing and/or sharing data.

One final thing.  What we try to advocate against is for blanket regulations.  So, some countries have chosen to put in their data protection laws that no data, regardless of what sector it becomes, can be stored or processed abroad.  That's what we think is very harmful.  You need to classify and determine those things depending on what type of data.  Very happy to share our research findings.  Thank you.

>> Just want to say, thank you for that intervention.  That was super useful.  We're seeing that play out in the Indian policy landscape as well, but I think we struggle to sort of identify what those segments for classification are, and we ran into that with the non‑personal data governance framework where it's sometimes hard to distinguish between especially also what non‑personal data is after it's deidentified, there is still harms and risks that make it sensitive to be shared externally.  As we think about classification, it also varies from jurisdiction to jurisdiction, so it's fascinating and we would love to hear about your study and see if we can make it part of the resources that we share with the participants today.  That would be fantastic.

>> AUDIENCE MEMBER:  Thank you very much for the presentation.  My question is related to besides the paradox of data sovereignty, there is the issue of even locally the data privacy issue versus the security or national security platform that refers to personal data issues so data can be taken.  If they do want to take it, they can take it.  There is not an issue of privacy.  Whereas the clear demarcation, even in not only Africa but in developing countries, even if U.S. in any security institutions can take private data behind the door.

So, are we ‑‑ where is the clear ‑‑ how can we resolve ‑‑ how may an individual be protecting the data, so how can we trust even the collected data of mine to be protected?  Thank you very much.

>> I want to start out by saying there is not a great answer to this one because I think we're dealing with it as we ‑‑ I mean this this moment, in fact.  We've seen it most recently with sort of the Indian privacy regulations that have been developed as we speak in some ways where there is this tension between data for national security purposes and what backdoors the governments wants to sort of install in the legislations.  It makes us increasingly hard for us as citizens to have the collective power and ability to push against, you know, against at that.

I wonder whether there is room for us to think about, and I think this is kind of the work we've been doing around stewardship, to think about how we can interact with smaller intermediatory organizations.  I think that's where stewardship has come in.  So, I think dealing with the sort of policy‑level questions are definitely complex and I don't think we have a good answer.  But I think the context in which we see stewardship is really at more of an individual or collective level where we're engaging with organizations that probably have a lot of sensitive data on us, even if it's an NGO, right.  This is the reason we talked a lot about nonprofits collectives that are capturing a lot of our information.  Are there organizational principles that these smaller organizations, the data holder themselves can start to institute?  Even if we are left without a sort of privacy regulation or legislation working in our favor, can we advocate or can we have organizations advocating on our behalf.  We faced this in India where we don't have a privacy bill.  If we do have a bill, it's favoring one over the other.  So how do we read?  Data stewardship has been our workaround or way of thinking about a workaround at that level because we don't have the sort of necessary power in other cases to sort of respond against that, or to ‑‑ yeah, so we see stewards as sort of these intermediaries that can push back to some extent, but I think again this is an open discussion and we'll have to see how it evolves in our developing countries as well as in the developed world.  I think even with the GDPR, this is a bit of a concern today as well so we're not seeing that other parts of the world have figured it out as well.  Would you guys like to?

>> I'm sorry?  I think that the question also responds to why we ‑‑ why we have ‑‑ are studying data stewardship because we've recognized that people, different people have different needs.  And so when you ‑‑ when you convene and come up with communities and cooperations, you serve your needs.  You contribute your data and say that you want to ‑‑ I want this, whatever value ‑‑ whatever value you want from that to come.  So, it's like having like frameworks on a national level assumes that everyone wants to participate, it's kind of like facilitating forced participation.  And yeah, and doesn't consider the fact that different individuals have different needs, and so we want to start thinking about the fact that people have different needs, and that you know, if this one system doesn't serve your need and you want your data to be protected from there, you don't have ‑‑ you have the right to not participate.  I mean I think overall we don't have the framework or how it looks on a national level or continental level, or like protect shows for instance, where you may be having this, I don't know, this system or dataset that does have value or does contribute to some sort of like objective for community, but then it's like most of the value goes into this international organization.  It doesn't deal with that, but I think we want to also like really break it down towards like these little communities.  Yes, yes, it's data, but it's data for like a specific person, and we also have the same needs, and therefore it should look at it.  That's what I want to say.

>> AMRITA NANDA:  One more question?

>> AUDIENCE MEMBER:  Thank you.  I think data stewardship actually was raised earlier by different participants, and it has a different way of what is data stewardship from government or national perspective or organizational perspective or civil perspectives.  Mainly as to my understanding so far, responsibility for handling and controlling and sharing data from national or institutional or sectoral perspectives, that is a way data stewardship mainly is used.

Now, as I from your thought and case studies presented earlier.  It is from human perspective as an individual, or would an individual be a member or part of a community.  Now, the questions that raised in my mind is that, yes, governments may own some kind of data, organizations may own some kind of data.  And now there is a very basic issue.  Can an individual as a human being own data?  Individual data ownership, is there such a concept?

And then again, this kind of ownership at the community level, individuals belong to one or more individual communities, taking four or five cities show and they belong to different communities.  So as a community, is there a concept of data ownership?

And another perspective relating to this one is governments in a legal perspective, define the right to access data, the right to access data as an individual I have a right, a right to access my own data, my personal data and any data collected from my part by anyone from the government, that may belong to my property, the right.  So how to access it, using what kind of platform, and how do we share with others, I think that is the perspective we are trying to highlight in advocating data stewardship from the human perspective.  So, it's better to correlate and see the difference and different concepts that we're using in data governance and how it's related with data stewardship from different perspective, and how it differs from the other one, and also, it's better to link with also data ownership at individual and community level.  Thank you.

>> I'll keep it brief because I know we're almost at time.  On data ownership, it's always been a controversial way of thinking about data because likening something that is basically an extension of our human self to property, I think especially from the legal lens comes, and from the ethical lens also, comes with a few outcomes that I'm not sure most governments or broadly, you know, thinkers around data rights are ready to ‑‑ or would be in support of viewing data as property.

But one area where this thinking has in some way been applied or sort of tested is in data trusts, where the sort of common law conception of trust is transported to a data layer but even in that case, typically data trusts are envisioned as where the right ‑‑ where data rights are the sort of center hole of the trust and there is no actual ‑‑ there is no sort of signing off of data as a property, and I will also sort of caveat that by saying on our research on data trusts, which has been extensive so far, we find that's a model that only really works in common law countries which are very few in the world, and even then it reflects a more top‑down management of data, rather than something like the human participation and that.  It's the way we envision stewardship is not around ownership for exactly that reason, so we don't view it as property but we view it as something that requires collective negotiation, that requires meaningful participation.

Maybe ‑‑

>> I was just going to qualify that to say as well I think the ownership from an individual lens kind of precludes the ability for us to have collective value over data.  Data can't be just used by one person and then it's done.  I think we see data used by one person and also by another use possibly, for community use.  And also, many instances there is data that may not be attributed to one person, maybe attributed to the community and has value for multiple community members.

For instance, the data that generated by sensor in the backyard and deals with air pollution, doesn't just belong to you as access to a sensor, but may belong to the community or residential area in which you inhabit, so I think the reason we stayed away from the sort of data ownership lens or that lens is that it kind of ‑‑ it avoids the possibility of us having multiple or secondary uses of that data, whether it's for research, whether it's for innovation and I think there is also data, the value of it can remain kind of hidden unless we share it.  So, in some instances, yes, you absolutely want to keep data private and should have privacy over it.  That emerges in the data sovereignty movement and that's another thread actually very well tied to data stewardship and in that context, has certain principles where the indigenous community has complete control over whether they want to share data or not because any data that's shared may be used in ways that could harm that community or set of individuals that belong to the community.  So, there are some instances where data ownership comes up, but I think it's tricky especially when you can't separate data points or individual data points from a large are set, for instance, and fails to consider the larger collective value around data, the secondary reuse of that data, so I think it definitely comes up in the discussions but we stay away from it because it's super loaded as a term conceptually and how it's been applied.

The last thing I'll say is that it tends to be, sort of immediate thread that comes from data ownership is okay, how can you then monetize the data and it becomes like easy popline to the monetization question and I think there may be value around data monetization or not, we're not saying that's not the case, but a lot of what we highlight is also the non‑economic use of the data for individuals and communities as well that could be collective in nature as well, so I think that's kind of just the approach that we've been thinking about a little bit more.

>> I was going to add the point on monetization but our colleague Roland who wasn't able to join us completely is on the call now.  Maybe we can hand it over to him for some closing comments and just to close out the session, and we can of course take this offline where everyone else who has participated.

>> ROLAND BANGA:  Thank you so much to my colleagues for a wonderful presentation.  I wasn't prepared for the closing remarks, but just to speak more broadly about what is the work we've done at a continental level, regarding putting together a data policy framework, which really seeks to ensure the ever‑increasing production and use of data across the continent, benefits all Africans, considering we know issues around data governance as well as data stewardship and the policy framework looks at the essential infrastructure needed for realization of value from data and assisting of creating a safe and trustworthy digital environment that supports the development of our sustainable and inclusive digital economy and society.

So, I would like to thank again my colleagues as well as the participants for the interaction that you've had.  We will share the resources after this presentation that really provide links to the work that we have done around data stewardship and encourage further interaction and any questions or comments that you might have.  Thank you very much again.

>> Thank you.

>> (Speaking off mic).


>> All right.  We mentioned resources as did Roland, so we mentioned briefly we're working on the playbooks.  Next slide.

So, we have a playbook that we've been working on, so we talked about shared responsibilities, the fact of data ecosystem support and multiple layers need to be considered if stewardship were to be put into practice.  So, Aapti has been putting together a playbook to deal with strategies and action items that everyday stakeholders can do.  We have four plays we put together.  One that you can do to ensure participation.  One what steward can do at funding and sustainability objective, main entities that are not monetizing data or abstracting data, how can they be responsible entities funded and sustainably funded?

The other looks at actual technical mechanisms around participation, so we talked about a sociological level, data justice level, and how to we start to infuse this into the technical architecture is one of the plays that we look at.

Then the last is we look at sort of public sector institutional mechanisms, so many of us in the room today are think being how to frame this at an institutional level, at a policy level, at a state level, so how do we start to think about participation at scale and at ‑‑ and by sort of public sector official, what needs to happen to sort of push the envelope further down the line in so we wanted to make this playbook available and accessible to you, and we would love to do so if you would be able to share your emails with us or if you have cards, and just sort of keep this conversation going.  This is something that is very much in the works, and we hope to use the considerations from the room today to inform also what we have put down in the playbook, and keep this as a dynamic resource.  We are a public research organization, so we want to keep this as accessible as possible.  If you don't mind sharing your details with us, we'll take it down or sharing your card with us, we'll email you with a copy of the playbook as well as deck of resources that we pointed to throughout the presentation.  I want to thank Roland who couldn't come in earlier, thank you for your patience today.  This was really fascinating.  Thank you for all of your questions.  Please do stay in touch because we're keen to keep the conversation going.  We didn't have all the answers, but I think this is something we can collaboratively shape answers together.  Thank you once again and appreciate all of you being here today.