IGF 2019 – Day 2 – Estrel Saal B – DC on Innovative Approaches to Connecting the Unconnected

The following are the outputs of the real-time captioning taken during the Fourteenth Annual Meeting of the Internet Governance Forum (IGF) in Berlin, Germany, from 25 to 29 November 2019. 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 to understanding the proceedings at the event, but should not be treated as an authoritative record. 



>> MODERATOR: Okay. Good morning. Good morning. Let's get started. This session is the Dynamic Coalition on Innovative Approaches to Connecting the Unconnected. I'm Michael Kende, one of the co-conveners. Great pleasure to be here.

A few quick words of introduction, every year we hear about the numbers of connected and it's been a -- with the ITU and it's been approaching 50% for a few years. And now it's past 50%. We hear the reminder that 50% is great but that leaves the other 50%. We don't hear the reminder that the growth rate is slowing down.

Every year, the growth is going slower and slower. It's early in the adoption cycle for it to be slowing down. Hence, the need to innovative approaches for connecting the unconnected. So, we've been doing this now, this is our fourth time at the Internet Governance Forum. The first in 2016 talked about the first innovative efforts to Garth dear at an on innovative approaches that Christopher Yoo will talk about in a moment. But that's really to try to be more systematic about what's working and what's not working on both the supply and demand side, to try to figure out ways of overcoming this slowing growth and keep the -- keep the momentum going, connecting everyone.

Then we invited a number of grassroots organizations, some of the ones that have been studied in these case studies. And it was fascinating discussions about the efforts that people had been taking in different countries around the world to connect the unconnected.

And last year, Christopher and the team introduced cost data from a number of the case studies to start to synthesize them and see what the learnings were from all of these case studies.

So, this year, building on all of that, there's going to be three topics discussed based on this initiative One World Connected that Christopher will talk about. And these three are just an overall cost analysis, looking at the case studies and the data that's been gathered on the demand and supply side, and then looking at two areas of particular concern, efforts around gender and efforts to build up digital literacy.

So, what we're going to do is there will be three presenters from the coalition that will be presenting on each of those three topics. Each of them will present for about ten minutes followed by a commentator talking for about ten minutes, which will leave us half an hour for what no doubt will be an excellent discussion, questions and answers among you.

So, without further ado, it's a great pleasure to introduce the driving force, the engine behind all of this, Christopher Yoo, a professor at the University of Pennsylvania. Christopher?

>> CHRISTOPHER YOO: Thank you, Michael. Thank you for coming this morning. Seeing so many old friends and some new ones. So, it's a pleasure to reconnect with people.

Getting in the substance, that's what we're all here to do. If you would, hit the slides, please?

So, One World Connected as Michael said something we've been doing several years. The mission we have is to create -- is driven by exactly what Michael said to try to deal with the slowing rate of adoption when only half of the people in the world are connected. And what we discovered was there was a lack of real -- many people were trying new things but no data being collected about what was working and what was not.

And the biggest problem for me is that what many ministers were being -- were hearing was sales pitches from equipment vendors or particular business interests, or people falling in love with a technology and advocating for it, and in ways that weren't necessarily data driven. My personal conviction is that there's no one magic technology, almost all of these deployments will involve different ones in different situations.

But any way, people would lock into one idea without actually having any validity. So, we actually surveyed 14, I think, major data bases to call to identify every innovative deployment that we could find, a way to -- very innovative program to attach, connect more people to the internet. In fact, we made a decision to identify all of them, not just the successful ones, which are the ones that get covered in the press. And we now have a data base created to do that. Then we contacted all of those -- we have 1,029, I believe, at this time. We have contacted them all and asked them to do case study interviews, we have 120. We structured the interviews?

A way to try to develop cost and estimate of cost performance and revenue performance to try to understand which ones were sustainable, likely to be sustainable. And that really to try to create an information-base to drive the future.

The next direction is in the process of fielding, answering a different question. Not just how to connect people, but what does that do for you in terms of sustainable development goals, healthcare, education, and particularly economic know. But we're using the gold standard of research, controlled trials. We're fortunate to have partners around the world. Happy to share what we're doing with that as of now. The next phase for us is disseminating the findings to try to share with interested audiences what we've learned.

Let me give you a brief introduction to the case studies and the data base. The data base itself is 1,029 entries that span 151 countries, the largest share in Africa and in Asia as you would imagine. But interestingly, there are a number of case studies in North America and Europe which we don't often think of as meeting part of the digital divide. You can think of the tribal lands in the U.S. or Scottish Highlands in Europe remain as difficult to connect as other parts of the world.

We found a division of the supply side and demand-side initiatives. Supply side is infrastructure. Demand side is a reality I think most people in this room know. We used -- there's a line from a movie, "if you build it, they will come."  Many surveys have confirmed that is not true. You need to do digital literacy education, relevance education, E-government, other services. No one buys an interconnection for the sake of having one, it's to get the services through it. We have a higher demand of demand-side interventions than supply-side. I surveyed 151 countries in the data base and many people found it to be a useful research tool.

The case studies are 151 spanning 151 countries, the geographic picture is similar as with the demand and supply side. We have a distribution of urban and rural projects and different geographic scopes, national, local, regional, and international. Happy to talk further about that if you like. I'll apologize in advance, I will not be able to stay in the question and answer period because of another obligation to speak in another session. But the rest of the team can answer any questions that you might have, because they've done the lion's share of the work of actually generating the case studies.

So, the first most surprising thing I will tell you is that 62% of the projects we surveyed in the -- in the case studies have no revenue whatsoever. And this is a problem because their funding sources are often from grants or corporate social responsibility. And the reality is, when those sources stop funding these projects, they will probably cease to exist. All of them are attempting to transition to paid models, but the problem is all of the attempts to do so have been largely problematic and very difficult to execute. Even more importantly is the problem seems much more acute for demand side initiatives than supply side initiatives. Still a problem on the supply side, nearly 40% do not have revenue. But on the demand side, nearly 75%, nearly three-quarters do not. What this tells you is if you're doing a digital literacy program, you can't just do it for basically free. If you have no means of then trying to gain some benefits from the people who preceded you, the idea that you could turn this into a sustainable program become s much harder to do.

If you look at the distribution of how the types of different funding sources, what you discover is roughly 80% of grant funded, government funded, or corporate social responsibility funded projects have no revenue model. The subscription models and community network models are better. This is an interesting change. We've advocated to grant-making agencies to change their approach, to start to look for sustainable business models, to fund projects that have revenue. To validate possibilities for the future, and we're actually gratified to see many of them have changed their policies as a result.

What's interesting, we've talked more about different aspects of the program, happy to do that as in past years, happy to continue that discussion. But I wanted to share with you some preliminary results about cost effectiveness. The key for us is to collect data, which is rarely done, but to do it also in a way that permits cross project comparisons. We were told by a number of people at international organizations that they appreciate this, it's needed because politically, they cannot pick winners and losers among projects and academic-based research project can.

So, what I will tell you is we set up a metric through our structured case study interviews that allows a sort of cross project comparisons. We have a number of financial statements shared with us by the projects. We have our structured interviews and we asked all of the projects to fill out templates about what their deployment was. The revenue side is based on what they were charging if they had revenue at all to make an overall estimate of their total revenue.

So, to share you some preliminary findings. What's interesting is many people worry about the cost of building the networks, the capital expenditures. The biggest headline is the operating expenditures end up being far more important than the capital expenditures.

If you cannot operate on a profitable basis, you will actually borrow money and then continue to lose money with every year you operate. And so, I think many people focus too much attention on building networks and not enough on running them. And particularly we often focus on the last mile. One of the biggest costs is the backlog costs themselves.

Most of the deployments, the largest numbers we have are two technologies, TV white spaces and Wi-Fi deployments, often paid Wi-Fi deployments. The cost of TV white space s tends to be higher than Wi-Fi, primarily, one large driver is the different in spectrum costs. So, there's a policy interlay here where the spectrum policy of the governing institution has a large influence over which models can be successful and which ones are not.

The second driver is the expense of TV white space's radios. I will note three of the four white space s are pilots. TV a new technology, radio may go down. One deployment is one in India where a university is trying to create their own equipment. The costs are lower because they're doing their own work and they don't have to pay a commercial deployment, but you start to see the very innovative ideas.

You see the Wi-Fi deployments are much more widely varying in costs, they tend to be lower, but they tend to be much more variable, depending on the nature of the deployment.

A couple of things about costs, what we discovered is just comparing per-user costs can be misleading. For example, the projects that have revenue tend to have higher per user costs. Why? They're investing in higher-end services so you have to look at what's being delivered compared to the costs instead of just doing simple comparisons, you have to consider how large the deployments are, how many they're reaching to make estimates of how these are going before.

Another issue that came up in previous discussions at the IGF is whether deployments of connectivity to anchor institutions, these are community centers, schools, libraries, is an adequate substitute for connectivity. Many people believe that you need to give every person their own device. What we discovered is, that's a great aspiration, but anchor institution distribution can reach many, many more people. And it -- it is a good option when in person-to-person -- every person having their own connectivity is not an option. That's quite interesting.

Our preliminary -- well what we discovered is roughly half of the ones that have revenue, we talked about the supply side, it's about 60%. Have revenue. Of those, the first estimate, we're looking at the model, two of them only appear to be breaking even. So there does seem to be a revenue problem among the ones that do have revenue. There is wide variance on parity. Different countries are in different parts of socioeconomic development.

The last and many ways the most exciting part of the work that's developed now is going beyond connectivity to measure the impact of connectivity. We're in the process of fielding three controlled trial, one in Rwanda, one in Nepal, to see the impact of not just connectivity, but what it gets you in terms of economic development, healthcare and education.

We actually believe this is a critical component to the dialogue because it's easy to get a communications minister excited about connecting more people to the internet. But if you want to mobilize a health minister, an education minister, a finance minister, and a Prime Minister, they have to ask you, what does it get us as a country? And the health ministers we spoke to who are even committed to the idea asked me, should I give this 5% of my budget? 10% of my budget? In a world where I have many other economies to deliver. They need and in fact investment finance committee says this sort of data is critical. We went to this literature thinking we would draw on existing models and we found out there are no existing models. It's never been done. So, there are many retrospective models but a true control trial has not been done. But the one in Rwanda is based on the deployment of the company that's building 60 sites. We randomized across the 30 to try to get an estimate. And, in fact, we're getting -- we're working with the government to get data on economic development.

Vanuatu is a deployment using the connectivity to do remote diagnosis. It's in the second largest island, and there's no hospitals on the island, so this is really to determine so this is when the resident should take the high cost of travelling to the island. The two hospitals they can travel to are on separate islands specializing in different things. It becomes a critical -- early information about where to go become s critical. We were fortunate to have the chief and the head of this project to attend the IGF two years ago. And in Nepal, there's a health deployment looking at giving a pregnant mother apps and connectivity to improve newborn health outcomes. We've done the baseline estimates for an all of the projects. We're getting prepared to go back and do the retrospective estimates just to understand how this goes. And some of the countries the government has changed so we're in the process of renegotiating or understanding with the new ministers to get access to the data, but we are optimistic that those will all be completed properly.

So, what comes next. We're continuing to work on cost modelling. In addition, there's interesting work we're doing if you see four lines down on fixed wireless projects, four in the U.S. have given us complete data where we can do more detailed analysis. Developing sub reports by region and academic papers around topic areas within this demand side, gender, M health apps, digital skills training, the role of partnerships, and challenges. One of the more interesting sources is we had all of the projects report their most significant challenges and trying to figure out differences across different types of deployments to figure out what areas -- urban versus rural, national versus local, all of the differences in how they play out.

We're doing more. We're continuing with the controlled trial s with hopes to expand them to Latin America and health centers in Latin America and another one in another country in Latin America. And the last there is we could use the help of the people there in the community. Our goal is to try to find venues for disseminating the findings in the next year to interested bodies.

If you have suggestions on places where that can happen, I really welcome your input, thank you very much.

>> MICHAEL KENDE: Thank you. That highlights the value of gathering data to help to guide the impacts and how money should be raised and spent on these kinds of initiatives.

I would like to turn it to the commentator, Jane Kauffman from the internet society, Jane?

>> Jane Kauffman: Thank you very much. Welcome, all. It's a pleasure to be here. We're focused very much on infrastructure development. We have been working with community networks for over 13, 14 years. And as Professor Yoo has said, there's some question in some people's minds about the type of longevity and sustainability in community networks. There is longevity and sustainability. I would say we have to think about this paradigm shift perspective.

These are startups, in the world of economic development, if you're a startup, you might get startup vending from venture capitalists. If some of us, APC, the internet society, or another organization, the fund if you know Georgia, putting in money, small seed money to build up the networks. These are community-built infrastructure. A lot of people in the communities don't have bank accounts, they don't understand business plans. So, we talk about sustainability plan s instead because sometimes it's scary when you're talking about a business plan.

So, I believe over time you will see a shift in that economic development model in the remote and urban underserved areas. We've seen viability in New York City in New York City, believe it or not, that's a community network being built out, Wi-Fi-based.

There's a community in Georgia. We're seeing return there. It's not just the communications ministries that are involved and not just local business but also international organizations coming in to say what can we do to help you, not that we know better. If you come in saying you have the solution, you don't. You don't know the organizations, the countries, even that region. So, two weeks ago, we were in the republic of Georgia in Tbilisi. We're seeing new interest in the community itself. So, we have to look at impact from a different perspective. It's not just about are they immediately bringing in revenue. It's what is the community-based model for sustainability? You have to start in rural areas with a shift in mindset in coming about. So, we've seen 50% increase in air B&B. Reservations in this high-mountain village in the middle of Tusheti, 4,000 meters above sea level. There are people connecting with other people in the capital. They're bringing revenue back to this area and it's rejuvenating tourism. That's good/bad in terms of bringing more people to the area. But it's giving the Georgian government a paradigm shift in change. If we stick with the old model of how companies build network, we will never connect people.

Not popular when I say this, some companies -- I used to work for a big company years ago, I know that through change and licensing, Universal service policies, telco mindset in the regulatory bodies has to change. In communities in 5 and under, major companies can't get a return on their investment. So, if that is the case, we have to change the shift, the paradigm.

So, we come in with smaller network, sustainability training, with technical local training. So, from my perspective, after seeing what's happened whether it's in sub Saharan Africa, Latin America, five to six governments have changed spectrum policies and looking at new licensing. So, I believe we have to shift our focus on what we collect as far as data goes. On economic development and how it's measured.

I'm not a expert. But I feel it and I know it having done it for 27 years.

This, for me, is real in the sense of rural development. And it is empowering the local communities, though, because you can't go in and say this is good for you. That's ridiculous. A lot of people will do that. So, we've got to figure out a way to work at the local level. It's not easy. It takes time. We've seen that with internet exchange points. It takes seven years in some countries just to get to the point where the regulator and the incumbent can come together and talk about a change in the policies, right? So seven year is a while. And you is eve got to hang in there.

From a development perspective, from the traditional agency's perspective, I couldn't agree with you more. We have to shift the billions of dollars or millions coming in and people will say, why do you think you should give them only 10,000 or 5,000. I'm like, how do I know what they need on that level. You do have to do an assessment. But if you're actually -- we've had community networks say to us, please don't give us that money. We'd rather have training. Because money at times can corrupt the local system and create a strange competition that's not sustainable among local people. You don't want to see that.

There are people in this room more expert than I am. But I get worried about agencies that say -- it's weird for me -- how can a development agency in this day and age with all of the technology say they can't give out small amounts of money. We hear this philosophy that it costs the same amount of must be to give out $2 billion than it does -- from the infrastructure inside of an organization to give out $10,000? Seriously? Can't they change that? I think they can. So, many of you work with development agencies, I used to work on projects in the field. You've got to take great care with the community that you're working with. You don't want to corrupt that community. You have to think about the change and you really -- I think we can do this collaboratively. It's not just internet society or the university. We've all got to come together and almost put these factors together and say we're seeing this change, we see these factor, how can we work with you to change that? And is it a holding organization that can take that money? Foundations, maybe, gently push it out. You have to take the temperature of the community so you're not creating difficult change or corruption in the community or local conflict. Because we don't know if that's -- at certain levels. There you go.

>> CHRISTOPHER YOO: Thank you, Jane. And I want to emphasize things that you're saying that are really critical. It's that I do think we need new approaches. What's exciting is you mentioned community networking, community networks. And we have Carlos Miranda who's a leader who built network, insights with this. 25 of our case studies are of community networks. What's interesting is I see a sea change happen. The program I'm going to leave here from here to speak at is being convened by two trade associations of traditional operators. It's ETNO and OSEAT in Europe and South American. One of the topics is how do they relate to community networks and a maturing of their perspective to understand that community networks can reach communities in area where is other places can't seven. This is critical.

Another thing Jane emphasized that comes out clearly in our research is that capacity building is essential. You have to figure out many times a grant -- a grant making organization or a government stands up a build but doesn't create the capacity technically to keep it up. It doesn't stay up. As we know, all equipment needs to be updated, maintained, something in our research, sometimes the latest and greatest technology is not always the right thing. Sometimes it's the simplest and easiest to maintain and the most off the shelf that becomes critical.

So, understand the capacity of the organization and the community you're in and building those ties becomes absolutely critical. But I think, you know, exploring within these different worlds, different models that work and don't work, some of them -- there does have to be a way to think how can we sustain them in the future. Some could be sustaining government support. One of the most exciting ones, a paid Wi-Fi model on U.S. per month per user training a lot of people. That's nothing magic about that. You have to understand how that fits. But there are a lot of interesting ideas that many people didn't think were possible that now seems to be succeeding.

>> MICHAEL KENDE: Great. Thank you. Since Christopher has to go to the other event we can take some questions on the first phase of the -- the first part of the presentations, the costs and questions for either Christopher or Jane, please?

>> AUDIENCE MEMBER: I put about this figure that half the world is connected. For me, the affordability factor is key in the people that are connected. So, to my mind, actually, of the half that are connected, only three-quarters are barely connected is what I would put it. So, we need to take that into account that many of the places that are so-called connected, people don't have the funds to be able to use that connection effectively because it's on a high cost mobile network where there are charges.

The next question I have, have you avoided looking at consumer networks? I assume you have because there must be like 1,000 municipal networks in the U.S. alone doing this kind of thing already. That local authority provision of services is something that needs to be well known and well understood.

And my third observation is about the emphasis on the economics. This came out interestingly enough in the Colombian IGF is in the rural areas, the conclusion may be that it may never be fully economic, economically sustainable to connect those people in the remote areas where they're vastly populated. To look at the social benefits and to try to use the observation that the social benefits are really important for digital inclusion in the countries and to use that as some form of metric for driving government support for those areas. I would like to draw your attention to the pathways commission report last year which said connecting people in poverty is predominantly a matter of affordability, but the business as usual approach, setting prices to recover infrastructure investment will never be affordable for the poorest in society. Thank you.

>> Thank you, Mike. Always a pleasure to interact with you. A couple of thoughts. We did not do -- within the study, we made a choice not to look at municipal networks. There are not thousands of them in the U.S., it's probably in the low hundreds, if that maybe 1100 and change.

I'm doing the work -- pulled the statements of every fiber build in the U.S. I live in the city where they try to build municipal Wi-Fi. To be honest, the story of municipal broadband in the United States is not a happy one. They largely fail. Partly lack of political support, it's one mayor's pet project when the next mayor comes in, it's hard to create additional support for someone else's success. But also, what my early take on -- my -- the data isn't completely clear. The hardest part, though, is whether -- I'm trying to study whether they fail because of too high capital costs, too high operating costs, or insufficient revenue per user. My conclusion is leaning towards the last one. Because getting people to -- you're -- it's a constant marketing game, particularly if you have an area served by multiple providers, you're bombarded with ads. I think the cities are very good about running the networks. But it's a fulltime job, particularly when there are existing providers. They set up projection s that are far too optimistic. They think they're going get 70% adoption. Every provider tells you no one gets 70% of adoption. It doesn't happen. It's hard to get 30% in a competitive environment. You need to come up with some fancy marketing campaign all the time.

Too much emphasis in economic, I do sort of -- I don't want to oversell this, but I think the economics matter. I think there are more cost effective and less cost-effective ways to do it. Even if you're talking about a build that will not be sustainable that will rely on government funding, learning from the past about what's working and not working in terms of technologies, in terms of business models, I think is critical because in this day and age, even now the government has other needs it needs to fund. And the extent to which we can do it more effectively, so much the better.

We have some spectacularly well-ordered projects that failed miserably. And trying to learn what people were thinking, because these were smart, very well-meaning people. But to try to understand what led them to the point they were doing and understanding how that's -- what went wrong, we think is an important step. And to find out if they're going to rely on alternative models on subscription-based funding, governments need to make the commitment they need to find out what the size is and how to make that to make the builds stay up in the long term.

>> Jane Kauffman: Do you mind if I throw one thing in for you. I think there's a factor that's missing and ask you to take a look at it. The reason so many of the muni networks haven't survived because of the embedded historical rules and regulations. There was a policy in regulatory regime that is anti-muni, anti-community network. I'm from Maine, rural area. I have to go wave my phone up like this where my parents live in order to get a signal. And part of that is because of the old infrastructure and the old networks.

So, I'm thinking about doing what Steve had to do and see if I can work with my mother, used to be a councilman, she's a volunteer, by the way. It was her night job, to work with the town, to sustain the town.

This is local volunteerism. But it is sustainable. You'll find there's more demand and bankable demand in some rural areas. This is, of course, in a place that has -- well, there's poverty, a lot of poverty in down east Maine. But in any event, there are people who will afford connectivity. I got a ping this morning about young college students in the democratic republic of Congo for going through connectivity. Carlos has these stories as well. And I don't want to sensationalize that. But I want to say people make choices. And if you're choosing connectivity when you've got to make a choice between food and networks, that's really sad that we're at that point.

There is going to have to be some government support. Let's be honest, everything is subsidized in the old telco model too. There were great subsidies for the companies and incentives. That's where Universal service funding too. Let's think of different ways we can do that. But let's look at the policy regulatory or municipal rules, there are some networks that try to get off of the ground with modest funding but got just killed because of the local government saying no. They were swayed by the big companies. Now, I'm not about the company, because we're seeing a lot of the great work Telefonica is doing a great project in Peru. We're changing the mindset. And I would put out a modest proposal for the big companies if you're not getting a return on investment in communities of 5,000 and under, why aren't you creating smaller business units to go do that? No, this is not the solution for everywhere because some people want their own local sustainability and political governance. So, some of this is very much tied to local rules from whether it's tribal, sovereign. So, I would say from a modest sense, I think we've got a real problem with embedded historical architecture.

>> MICHAEL KENDE: We have remote participation, so the mic is essential.

>> AUDIENCE MEMBER: Hi. Hi, Nikola. I was lucky enough to work with Mike last year looking at the impact on community networks in gender and social impact. And one of the things that Mike brought up and Jane mentioned a little bit, but I want to crystallize a little bit in academic terms, if you like, that impact and paradigm s become a totality. So, one of the things that I frequently hear when I hear resilience is actually, I hear near liberalism. And, so, when we are looking at this -- this might be a bit of pushback on this idea of the gold standard of controlled studies, and how nice it is to populate all of the results on a one-page schematic that you can give to a policy maker. But it doesn't include all of the knock-on effects.

So, we have colleagues from the end to create a very good example that when we looked at community network s which are seriously bootstraps by funders, is it's a whole ecosystem that builds up. You have a community -- you have people who learned to build solar in the community networks that then go off and develop for their own franchises which work for the local network operator. And all of these things cannot be assessed in a snapshot one-year study. They can't be assessed by interviewing community people in networks. You actually need very deep studies of how these ecosystems emerge. And many of them have gone off to cities you don't know what happens to those people later.

So, getting back to my question is what can we actually do to unsettle this problem that makes an impact across efficacy of business models and we're only going to ever create the same story. Because we can't think of all of these many diverse impacts that cannot be captured in the -- in these types of studies. And what do we need to do to make this accessible to people who do want their one-page schematic? So, I hope that's not too hard pushback.

>> MICHAEL KENDE: Not at all. I hear what people are saying. I want to -- maybe part of the answer is that we think that our study will not provide all of the answers. We do think that understanding the costs of it is part of a -- a critical part of the question. And research is a long play. And even when we come up with one study, it's going to be one-point estimate of one intervention of one country where we won't actually even know how representative it is. But how you learn as you build that up over time.

I take seriously the notion that there are secondary effects from these deployments. The problem is, we don't even know what the primary effect is under things that we can measure. And that this study doesn't answer all of the questions that needs to be answered is not the right way to think about the value of a study.

First, is there any impact on healthcare from giving a maternal app to -- a health app to pregnant women? It's a great first order question. There's no doubt that may create collateral benefits. And those need to be measured. But that is not -- asking one study to answer all of the questions at the same time leaves us without any real evidence where you end up seeing people doing things based on very well thought out conjecture, very well-meaning conjecture. But as we know, some of the best thought-out plans don't work out the way we think and it's not until we get to the real world, we find out what works. I don't take it as a pushback at all. It's the modesty of how you build this kind of information base.

>> MICHAEL KENDE: Sorry. We'll move on. But we'd like to move to the next topic. One aspect in every country is that there are many people who could go on-line, who have networks available, whether the model is being studied or others. But haven't chosen to go on-line for a number of reasons, one of those reasons is digital literacy. Not knowing how to go on-line or how to make a meaningful -- make that access meaningful. To talk about that topic again from the University of Pennsylvania, Sharada Srinivasan, please go ahead.

>> SHARADA SRINIVASAN: Hello. If we could have the slides on the computer show up on the screen, that would be helpful.

I did want to start with the disclaimer, very fun. I recently started two months ago working at the global factors of the World Bank. And the results that I'm presenting here are not representative of those views. I maintain an academic fellow research at the University of Pennsylvania and all of the research I'll be talking about here are informed by the work I have done in the course of the last 2 1/2 years at Penn. That is -- this is what I'm going to talk about. My pressure -- presentation will be brief but hopefully meaningful.

Going to talk about the digital skills portion of our data base. I'm going to talk about what those skills really are. Because a lot of people claim they're doing training without a lot of synthesis on what kind of training are included and not. I'll talk about what we know from both our data as well as the literature. Lastly, I will go to some of the questions that need more research, answering, more thought from policy makers, some of whom might be in the room.

Quickly this, is not like it shall this is not very, very new. I just want to give you like that we have over 1,000 plus projects in our data base, 1029 as you mentioned, Professor. 150 projects we have case studies on, we have 24 projects that are out of school digital skills training across the different audiences, and we have 13 school-based digital projects. So totally that gets us 37 projects from 23 countries. And the distribution of that as you can see on the chart is like quite varied. We do have Latin America a little bit, Africa and south Asia. We don't have a lot on the Pacific training side. That's one of the drawbacks from our study.

Professor, you mentioned, as cross demand projects overall, skills training projects nearly three-fourths of our projects do not have revenue. Most of them are ad hoc, sustained -- funded or grant funded. Some are loosely supported by governments without having a systematic approach with dealing with it. And the skills trainings projects as with our overall projects data base are a big part of our case studies.  They are the biggest constituent of the different kinds of things that we studied. The community networks are the second biggest and social constrain s are the biggest.

What do our data show? Our data show most of our projects are driven and corporate social responsibility reliant. They don't have a reliable business model. We take the comment very seriously from early on that you have to not think about sustainability just in terms of the revenue, but in skills training projects, we need to think about how it will continue, how the curriculum will change, how they will be able to support newer training. And that can be from the different sources. But at this point, that kind of thinking does not seem to be appearing within the ad hoc projects we have studied. They also run with a grant and they don't have much way of follow on work from that.

The second thing is there's a wide variance of curriculum and pedagogy. What do we mean? We mean there's projects that are dealing with school children, primary, middle school, high school. Projects dealing with out of school youth, farmers, adult literacy campaigns in certain countries. But they don't have -- and we have deliveries anywhere between in person tele center training, in person skills training through the sustained program, app-based training, web portals. Language training.

So, there are different ways in which this is being done. No standard ways to decide what's the purpose in the particular context. It seems to be more based in the people proposing this grant saying we think this is -- and we think this will get us the grant to do it. But there isn't a clear -- most of the projects we studied across the digital skills they didn't have a needs assessment in the community and some of them reported it changed the way they delivered content after they got the money and went into the community and realized that the needs were different, needed to be in a different language, different level of basic/advanced study.

Last one, there wasn't much by way of measuring specific outcomes in terms of learning of digital skills. It's hard to see. Digital skills have a spectrum. Different organizations defined it in different ways, basic skills, intermediate skills, advanced skills. The EIU has a huge framework. Some people consider like social cognitive skills about part of the whole digital skills framing. Some think it's about machine learning, artificial intelligence coding-based skills. But we don't know what each of these skills are leading to in terms of outcomes so that the people that we're affecting.

This is one of the findings from our data that show -- that basically leads to a lot of questions on what we need to be thinking about. The other thing we also know is there's a huge variance in terms of cost. Professor Andre presented some of the cost savings on the supply side. But there's the initiative, what kind of grant, where it's from, how long it's for. What the duration of the program is. How many numbers of students it trains in person, facility-based, on-line based. There's a huge amount of variance. And the other thing that we see is that most -- like a lot of our projects are really based on volunteer-based funding, some form of community funding. And that's sometimes very beneficial as I would talk about in some of the later work but often begs the question, if we do not have this critical mass of volunteers and we're not able to galvanize it through the project that we start, how is it that we will continue to maintain that kind of skill set given that there's lack of -- there's a possibility of skills.

The literature has not focused on skills but it has focused on skill-based training like technology. We do know that provisioning just infrastructure, be it school, text, material, doesn't lead to any meaningful like without any other additional work leads to meaningful outcomes, right? Some of the digital skills project like the one laptop per child program don't really work, have not shown an impact in learning outcomes in the studies that have been done on it. Providing just computer hardware does not work. You customize it to the audience, you get more and more impact. That's the more recent strand of work that the literature shows. But it means you can't just provide infrastructure, skill s training in the abstract, but tie it to a curriculum and pedagogical training that might show a lot of results based on the context.

One of the open questions this, is where I'll end, we need more monitoring and evaluation on learning outcomes. We need to have a set of metrics on what digital skills learning outcomes look like. We need to regress and monitor them across projects in terms of what are we doing? It's not just that they have created a certification and gone out. Like does it mean to meaningful job, is it the kind of jobs they want? We need some work on what they are. We need evaluation on those. We need to think about the goals and skills training.

This is the criticism that came up out of the older computer literacy training model where we used to switch people how to switch on and off a system, a device, and teach them basic things on how to use software without telling them why using that software works. And some projects in their pedagogy has moved away from it saying we will not just teach you Microsoft Word, we'll teach you how to build a resume or how to use a mobile phone to get agricultural services instead of just saying this is how you operate a phone.

We need to mainstream skills in policy making. But right now, there are a lot of ad hoc approaches but country-wide skill strategies in the countries we're studying in the global south do not have a systematic approach. And this obviously require s to work in the education system but also in projects outside of the approaches. Neither of those in the project s we were studying had a coherent framework so there were loose projects that were doing their own thing, but we didn't have an idea of how it fit together with the skills that country needs or that country is trying to advance in, and how digital skills training, in its huge spectrum, can help that. I'll stop here.

>> MICHAEL KENDE: Great. Thank you very much. Commenting on the Carlos more ray knows from APC.

>> CARLOS MORENO: Thank you. I'll speak from my own experience and I hope that all of the people in the room that might have more experience can also provide that input. Thank you very much for the invitation to come in. Around the studies, the first question looking at the presentation was related to the type of studies that would include it and how maybe those that are found online and funded by big funders are those that were identified. I wonder if how many of those don't receive that funding or the different marketing approach could be included around they could be included around many of the people I know in South Africa in particular learn from each other. They don't learn from the program -- I totally agree with you. I have some experiences around hardware that was develop in the schools and high schools that when we visited them to provide connectivity that Professor Yoo was mentioning, they were in boxes. Literal ly in boxes because there was not a program for how to deal with the confidence of the teachers, about how to -- how to --

I happened to -- my mother happened to be a teacher that went through the whole digital evolution in Spain at a time that the students knew more than her at the moment she had to teach them how to provide the -- or how to use technology to be included in the development that you were mentioning, right? The pedagogical improvement.

I think those questions are not taken into consideration when these programs are done, what is the psychology of the education? What is the confidence of the teacher to -- to be probably ran over by this. And what does it mean in places where education is based in a very hierarchal power relation between the students and the teacher, right?

So, I don't know how you can incorporate it in the study. But I think those are some of the thoughts that you might attach on the failure of some of the projects considering the social issues around -- the social issues around digital skills. Because as I was saying, many people learn from each other, learn from friends, learn from someone showing them something on the device, on the computers. Whoever has the device and the computers. There's a touch on what Mike was saying, there's a whole thing about the affordability playing a role on specifically not with the skills per se, how to use the device, but what do you get out in using the device. An there's the concept that's called general activity that is you don't know what you don't know until you know it and the internet definitely allows you to -- to learn things that you didn't know you were interested in, and takes you to new skills and takes you to new ways of engaging with that with your peers and the people around it. And the component of the skills training that might not come from big scale.

So, for that plays a big thing, a big factor. People cannot experiment if people cannot get interested on what they would do with those skill s. You were saying, need assessment. Maybe I don't want to -- I don't want to know about spread sheets and presentations. I want to engage with fun ways to create videos or uploading videos or other things. Maybe look at content and how the whole digital skills can be -- and I think it is sometimes more -- I learn more than the school. So how that can be incorporated in the study I think it would be interesting.  We're seeing more and more programs.

Another thing I want to commend about, touching around the -- I mean, everything, some sort of funding. But the whole idea of the economics that Michael was mentioning before around how we are considering everything within a market economy, with everything that needs to be delivered, is delivered, needs to be paid for. Most of the things that I learned in many life, I learned from people that didn't -- for them. Digital skills or other, many place where is the DIJ stall skills are not there in place, there are other systems, other ways of sorting and exchanging that do not necessarily fit in the market economy. Maybe someone shows you how to use a computer in exchange of a meal. It's impossible to incorporate that into the funding and into the other things.

So, I'm going to leave it there. Michael told me to be under six minutes. I'm there now. So thank you.

>> MICHAEL KENDE: Thank you. I think we're trying to save up enough time for questions at the end. So we'll move to the final steps of speakers, the critical issue that the group has been studying. And it's very important for access, the gender gap, the gender issues, so, again, finally from the University of Pennsylvania, Muge Haseki, please?

>> MUGE HASEKI: Thank you. Can I get the presentation? Hello, I'm Muge Haseki and I'm a postdoctoral researcher at One World Connected project. I would like to talk about our work on gender for the next ten minutes. In discussions around gender, it's often binary, women face greater challenge s than men in accessing technology in low and no income countries. Not all women experience the same constraints due to political or spatial factors. And One World Connected project, one of the goals to understand the barriers to a woman's technology, access end use, through in depth case studies and help to build more effective policy and sustainable interventions.

Why is this important? Every report from ITU shows the gender gap is changing fast in the developing countries. Interventions and policies are not affected to be able to digital divide. Therefore, we need to devise more effective strategies to identify problems and develop impactful interventions. The existing literature and reports identified six main barriers to technology, access and use. These are infrastructure, financial constraints, digital literacy, privacy, and social context. But these are based on case studies or the analysis of the existing literature. We're contributing to this work in two different ways. So, we are analyzing the barriers to a woman's technology and we're looking at the intersections of multiple demographic factors such as geography and age and our goal is to design policy and impact more sustainable interventions.

We have collected around 100 initiatives that aim to improve gender access and use of technology-based technologies. And in our data base, the majority are from Africa and Asia, these are the top two reasons with the largest gap in digital divides. And we generated 15 case studies from 12 countries in Africa, Asia, and Latin America. And these women-centered projects focused on literacy, health, and agriculture. So, some of the preliminary findings from our work. Women faced unique challenges in access and use to multiple and intersecting factors. Therefore, relaxing one barrier may not improve the end use. And the multiple barriers simultaneously can help to design more effective and sustainable interventions. I'll give you a few examples, some of the existing projects succeed or failed those intersecting barriers in their programs.

Here is one example. This photo shows that what happens to rural areas with poor infrastructure during heavy rain. But older men can still the mobile. It's often more difficult for a woman to be mobile and go to close the internet access points. It becomes more difficult for women with responsibilities for children and the elderly. And we have some case studies that shows in certain geographies such as rural Bangladesh, there are additional barriers such that women face resistance from the local community to travel by bicycle. There are social norms around riding bicycles for women. Even when women have mobile phones in rural India, they will not use them around their husbands due to status quo.

This becomes more challenging for women when they share mobile homes with their husbands and need privacy while searching for information on the productive health which is taboo in certain geographies. So, a community network in rural India created a woman on the access spots to allow woman to feel more comfortable using the forms to search for information.

In Ghana, families do not think learning computer skills are relevant. But they face multiple barriers compared to same age boys because there's a debate around when girls are old enough to use mobile phones.

You also look at the scalable projects different from those that are not and to what extent they add to those intersecting barriers. Here are four mobile phone-based services that we examined that aimed to improve the ma material and newborn health. And they scale in different ways nationally or regional or locally and reached a different demand of users with the intervention.

Some of the findings, allowing the woman to select the time of the day for the messages increases the chances of message received and read by the target users because of women roles and responsibilities. They're not able to check the phones or attend to the services offered to them because of these issues. And because women are not often decision makers in certain geographies, including gatekeepers to the intervention is essential for sustained behavior change. In certain regions in Africa and southeast Asia, there are more spoken languages. Therefore, creating local content and mostly spoken local languages is essential to reach the majority of potential users in a country.

And all together, developing interventions that address multiple access barriers, at the same time can increase the user examples. How skilled projects have the intersecting barriers in their programs.

Here's an example, have a scaled project John in Bangladesh is addressing intercepting barriers. Bangladesh and Nepal, pregnant women live with their husbands and mothers-in-law who are often the decision maker. But until Bangladesh target gatekeepers by sending them weekly health messages to change their health behavior which doesn't exist in Nepal which did not scale as much. HIV is an important problem for those in Burkina Faso that encourages them to reach out for advice because it's taboo. To access this intersecting barrier, mom collects anonymous messaging to the help desk, which, again, is not available in Burkina Faso which only scaled like locally compared to connect in South Africa.

You can see these three projects, you look at the number of languages and dialects available in the country and the local languages available in their intervention, and you compare it with the reach, you can see it's creating local content in the spoken local languages is essential to reach the majority of potential users in a country.

To conclude, some of the key takeaways from the research, face key challenges due to multiple intersecting factors, therefor, relaxing just one barrier may not improve the access or be sufficient to -- for the scalability of the interventions. Therefore, addressing multiple barriers simultaneously can help to design more effective and sustainable interventions.

>> MICHAEL KENDE: Thank you for laying out the challenges and some of the ways to get around it.

For some comments, Claire from the GSMA. Claire?

>> CLAIRE: Thank you so much. So, I had the assistive program. I'm going to speak about gender and in the basis of my comment. So we do a lot of -- we have surveys that are across 18 low income countries where we talked to individuals about their access and barriers in use. So, I'm going to share some of the findings from that, which I think reaffirms some of the points you made. But we worked also with local operators to help them design strategies to reach women with mobile internet and services. Working with them on the ground on projects.

So, just based on the sort of data that we have, it's important to unpack kind of the gender gaps, and I think it's to Mike's point, what do we mean by access? So, if you think about the numbers, 80% of the women in low- and middle-income countries have a phone that sounds impressive but they're not. There's big gender gaps. Also, first of all, gender gaps based on location. Significant gender gaps in south Asia, rural locations, so, there are a lot of gaps in ownership. And those who don't own a phone are the ones who probably would benefit from it most, most for this to weigh to achieve it. There are basic gaps around ownership for those who need it.

Beyond that, once you own a phone, it's not enough. You need to be able to use in it in a meaningful way. If you even own a phone, you uh want to go to the internet, you uh have to be aware of the internet. A huge gender gap, a bigger gender gap when it comes to awareness of the internet. Once you're aware, you need to be able to get on to then start to adopt and use it. Again, gender gaps grow. So there's an ownership gap, it grows around awareness. A bigger gap around internet usage.

Once you are using it, we're seeing in our data that women are using it for less number of services and less intensively than men. So, again, I think it's important to understand the gaps and where to target. The biggest barrier is affordability. And around the internet, the biggest barrier is literacy and digital skills to get on.

The work and evaluation of the initiatives that operators are doing highlights what Muge was saying is that there's a range of barriers that need to be tackled. And it needs to be tackled holistically. You can't -- and these kind of barrier s, some that didn't come out as strongly, affordability is a big one, hand set, digital skill, obviously a big one. Relevant content, a big one. But safety is also a big barrier. So, tackling. And it's stopping women from going on-line, so we need to tackle those barriers as well as a relevant content, yeah, accessibility to the services that you touched on.

So, if we're going to address the gender gap, we need to take a holistic approach to barriers, we need to acknowledge and address social norms, not just address the barriers, we need to understand what are the aspirations of the woman that also came up in the comments in the digital literacy that is not just about pushing and addressing barrier, one needs to understand what are the wants and needs of those you're trying to reach. My comments are, it's really important to understand the data. Really, we need to be data and evidence-driven. There's no one size fits all. We need to take a holistic approach. And we need to have clear targets of what we're trying to achieve.

Again, our data shows that the gender gap in ownership is -- while access is increasing, it's not -- the gender gap is not decreasing. So, we do need to take a targeted intervention. One final comment on that is that I know you were saying that you look ed at gender-specific initiatives, what we're finding in the work that we're doing with the -- with the mobile operators is not always about having a woman-specific product or service. It's about considering gender in the delivery of your services. So, thinking about the marketing, the distribution to the marketing -- are you marketing to women, do you have female agents in your networks?

How are you thinking about gender across your existing products, not necessarily about -- not necessarily launching specific woman services, but having that kind of approach and thinking about it holistically and thinking about the barriers and the needs of it when your targeting is important. And, again, when we've evaluated initiatives, those that just seek to tackle one barrier, it makes a difference but it's not what changes -- changes. It's when initiatives are having all of the barriers and taking a holistic approach that we see changes are happening. Our data similarly reflects what you're showing, I would suggest this is sort of a safety issue is something that didn't come across in your case. But we're seeing especially in some markets the biggest issue to be considered.

>> MICHAEL KENDE: Great. We have 15 minutes left. Take a question, maybe three at once. Since my back is turned here, I'll start behind me. Please?

>> Thanks for being patient with me. I was afraid to struggle with the machine there. I'm Jennifer Bouchier, from R&W media, based in the Netherlands, NGO. We build digital communities in places where human rights are under a threat. My career in gender equality and product development work.

What we found in our work is when you do get women connected, which is quite, quite low indeed, you still have to pull them out. One of the ways we found success is gauging them with women specific content. Not talking about backlogs and makeup -- sorry, my voice is a bit gravelly. We measured it in terms of, for example, Facebook participation and engagement. We have similar rates of participation in Yemen and Libya. But we have succeeded in having higher levels of engagement than the Facebook average. What we've done is a simple thing. We talked about workplace issues. We talked about getting a job as a woman. What are the challenges?

What are the successes? And really women have engaged. A woman blogger, one of the first in Libya, a young woman who was be putting out videos on the interest of issues to women. It's a small experience. It's not from a controlled trial. But I thought it was interesting to share this kind of result.

>> MICHAEL KENDE: Great, thanks. Please?

>> AUDIENCE MEMBER: Hello, I'm Carson from Tanzania. Would like to speak from the perspective from the youth and from a developing country. I come from communities which are really  agrarian substance base. People live under $1. Which means you have to modernize to go to the core of the city to get connected. The problems we have, the first one is we let the infrastructure -- we lack the infrastructure, there's no electricity. And another thing, people can't afford the equipment to get connected. And people get as -- as you said, you cannot quite substitute food or the basic need and internet. As a young person born in the annual, this is how I interact and learn and got the opportunity to be here.

So how can we kind of create alternative means first to fund this connectivity. And the other way of making it more relevant to our community case. We do have a community network, or one. But the problem is that the charges are very expensive. And there's quite a barrier of entry when you want to discuss with legislators in its meaning. You know? It's not meaningful now. How do we -- the government itself, the leaders lack this digital literacy awareness so they can make the lows or create the mechanisms on how we conduct the community challenges.

I was thinking how can we create a situation that can benefit my case, thank you?

>> MICHAEL KENDE: Thank you, thank you for those observation s? One more question?

>> AUDIENCE MEMBER: Excuse me. Okay.

>> MICHAEL KENDE: We'll go there and --

>> AUDIENCE MEMBER: Thank you. So, my name Lucas henson from the Internet Society of Pakistan Chapter. I was just wondering with the respective studies, if you had any numbers on the projects that you studied that -- how many were run by women itself. I would add a little bit of background. The project, the community and the project in a rural village in Pakistan, and the first phase connectivity was provided to that village. In the second phase, provided a girl's high school with a remote education facility where teachers from the capital were teaching the girl s the standard in English, math, science, and computers.

Now, the project, the head of the school, the first question she asked was, who are those teachers? Is any of them a male teacher? And this is sort of, you know, opened a whole new thought process for us. We realize that we need to have female teachers who are going to teach those little girls. Because that was something that was very cultural sensitive to that village. Yes, the teachers were females. And what we found over the study was that girls were actually more interested in coming to the class and their grades improved.

Just to get a little background on this, do you have an example on this?

>> MICHAEL KENDE: Okay. Thank you. I think -- okay?

>> Thank you. Excited to hear about women and children. My name is Samuel in Kenya. I would like to hear more about what adaption the committee is giving to the people with disabilities or people with alternative abilities. Because we're taking care of the children and the women. Both initiatives. What they are in terms of internet application and internet department.

Thank you. I think there was one more, please? One brief question about the methodology of the skills study. One of two things, I think it's how to measure -- how you measure those digital skills. I mean on the functionalist perspective, I can't imagine this, but there's also something about the critical benefits of being in a digital ecosystem. How would you measure this? Somebody makes a difference between news and fake news? Something that is really critical in today's world. But how can this be perceived from research perspective? And the other question is, where was the time span of the measured outcome, because I think there are like short-term effects that you can measure quite well, but there's some effects if you talk about educational programs and digital skills that would maybe take years to show them on a broader range than maybe the things that we have foreseen in the first place. Thank you.

>> MICHAEL KENDE: I'll ask all of the panelists maybe just to take a turn? Sharada, you want to go first?

>> SHARADA SRINIVASAN: So, thank you very much, Michael. I wanted to respond in order to some of the points that were raised both by the discussions as well as by the room. And then you might add to that. The way we looked at additional skills programs really was programs as opposed to more loose ideas around network-based connectivity -- connecting with people who can then teach your skills. I agree entirely with Carlos that's not always how we learn. We often learn outside of traditional educational systems, especially when it comes to additional skills.

We were constrained in that take given we were not taking an approach to the way we study these things. We did make an effort to go outside of the more marketing digital skills programs to lesser known digital skills programs. It was based on you had to have some kind of structure to do so. That clearly does have some drawbacks. But the hope is for people who don't have access to informal social networks that do have the skill sets, I was one of those people. I came from a conservative family that does not have any background in college education -- like I do not have -- I'm the first generation graduate in my family, access to technology was mediated very much by the ability to have access to people through the technology and breaking that barrier in itself for some people or some places can be difficult. The idea that institutional programs can help in some of those instances, not all. That's the modesty of the contribution. But as a hope that there's some element of importance to it.

I agree entirely that there is a lack in terms of teacher confidence. This is shown in some of the qualitative work that we did on our control trials work. We both went to one -- school s. Very similar challenges. Computers, the PC program exists but it existed in boxes in a lot of rural schools. One part was electricity. Not being able to -- the other was the teachers were so trained to teach from the text and not so much to teach from the demo in the class. And that really changed how much they were willing to incorporate technology in their own learning. In their own teaching experiences. That varied widely. And so clearly that's an important thing we need to account for that isn't given as much attention in current program training.

Do we need assessment again is very important but doing so before we decide on the structure of the program or the more delivery is more so. That's not happening at this time.

The, the thing about -- so, now I want to respond a little bit to comments from the room that had to do with DIJ stall skills. I will go quickly. I take from the back how to measure outcomes. The study right now is cross sectional and we take self-reported work from the case studies itself. We're not doing work in measurement and monitoring. The evidence has been people choose how many people are connected but not necessarily how much that connectivity means to them either short term or long term, because that's the current trend. I agree entirely that needs to change. How so is under question, especially around critical skills, fake news, etc.

There is some literature and education that's looking at pedagogy and critical thinking. But that is not something we've focused on very much in my research. Time spent on our research project, three years and ongoing. But it's cross sectional, not a longitudinal study.

Next, what do we do for people with disabilities? We have projects for disabilities in our data base, but very limited. So, one of them is on teaching to people in braille libraries, to -- to people who don't have -- who have visual disempowerment and are not necessarily available to access materials. We do have case studies, Project Zero, something we're not a part of, collects innovative ways to access -- to approach people with disabilities and we included that as part of the studies we did, studies on overall data base building.

On gender. That's interesting, it’s dependent on context is what we're finding out. In some context, having a male mentor in a industry for women can be incredibly empowering and there have been studies that shows having the male mentor take them under their wing and introduce you to a technology-centric world can be empowering. In other ways, women are important to teach certain subjects and see role models. That's there. The last thing I wanted to say is that there are intercepted challenges with supply and demand. In response to Carson, on community networks facing spectrum challenges.

The idea is is look at capacity building and when you think about conceiving a community network, you need to think about the community around the digital skills training, that's one way to think about it. A lot of networks are doing so already. And it's not -- it's not going to solve all of the ISHL shoes, but the afford about, it's not going to solve. But we are going to have to relate to different levels. Level of the community, the level of the regulator, the level of the policy to be able to interact and create change across the board.

>> MODERATOR: We'll  try to get to everyone? Maybe just a TWEET? Claire?

>> CLAIRE: I know we're out of time.

>> The summit in Africa took place in Toloma a month ago. More than 140 people. More than 70 of them were Tanzanians. Last week, the community network -- Tanzania was formed. So, there was definitely a lot of people around in the country willing to do something and the new community networking in an gore are a was announced with the support of national geographic. The minister on communications was at the closing ceremony. He made several commitments around supporting community networks. And rural spaces in particular. So I think you are among the luckiest at this moment in Africa. Reach out to -- (indiscernible) together with ABC, we can put you in contact with people in Tanzania. Because there are things that are happening that's probably the hot spot of the continent.

>> MICHAEL KENDE: Thank you, Jane?

>> JANE KAUFFMAN: Thank you. I would simply say let's think about different paradigms right now, working with industry, community network, working with the government, working with the development agencies. We've got to step outside of the box from the traditional way of thinking both from the economics analysis, from the regulatory policy models, and it's no it saying that we're trying to radically change the way spectrum is allocated or do away with the current model. At all. But it's a way of looking at allowing for those exceptions to become the rule and not the exception.

>> SHARADA SRINIVASAN: We have booth number one. We have the case studies as well as some of the snapshot of our youth-based digital literacy programs so if you would like to get some of the handouts, stop by. If you have questions or comments, we'll be there during the rest of the day.

>> MICHAEL KENDE: Thank you very much. So all that remains is to thank the panelists for a fantastic and very interesting discussion.

[ Applause ]