Keynote remarks at the USAID Digital Development Forum, 9 March 2018.
With entrepreneurial drive, the right technological solutions and a supportive, enabling environment, what could digital development look like in 2028 for base-of-the-pyramid populations?
If there is a guidebook for speakers in the technology community somewhere, it offers a warning: “don’t accept an invitation to talk about how technology will have changed in 10 years.” In technology 10 years is a long time.
Despite having made this first mistake, I’m thrilled to be here with you and to see many dear colleagues at this great event. Thanks to USAID and the conference organizers for setting the trap — let’s see how it goes and whether I escape.
I’ve put these remarks up on Medium. So if Medium is still there in 2028, and if you remember, you can have a look and probably a good chuckle.
To begin, and to manage expectations in the predictions department, let’s look back before we look forward. My daughter, Calliope, turns 10 this weekend, so 10 years ago, March 2008, is an important marker for me.
Perhaps we could begin even earlier in the story. We could anchor the start of communication for development back as far as the 1960s, from which the great “diffusion of innovations” approaches, championed by Everett Rogers and still in use today, emerged.
By 2008, with new names like “ICT4D” and “development informatics,” our worldwide community of practitioners and researchers, policy-makers and technologists was enthusiastic about finding new ways to apply a confluence of increasingly powerful technologies to development challenges.
Perhaps that was the year some of you updated the custom dictionaries in your Windows Vista machine to capitalize One Laptop per Child or to accept the European spelling of telecentre (it sounds much better with an ‘re’).
Regardless of what was happening with PCs, the big shift to mobile was upon us by 2008. China and India had bigger shares of the four billion mobile subscriptions worldwide than the United States. In Bangladesh, the Grameen Village phone program was reinventing the village payphone. Robert Jensen had written the famous fish paper, and even the buzzword M4D had entered our lexicon.
We were also seeing hints of the promise of digital financial services in 2008. M-Pesa was only a year old, but had already enrolled more than four million users in Kenya, with similar models appearing around the world.
The reason I raise mobile money here is that it was an early sign that the promise of digital development, led by mobile, would be more than calls, messages, and even content. The impressive thing about mobile money was that it was code: the systematization of a set of interactions that gave a new group of people the opportunity to do things digitally that they couldn’t do or couldn’t do as well otherwise.
Other signs were there, too,but faint. 10 years ago, the first iPhone had just been introduced — still only 2G — and many of our social media giants were only toddlers. YouTube was three. Facebook was four and, although it had a hundred million users, it lagged Google’s Orkut in India and Brazil, and Mxit in South Africa. WeChat and WhatsApp were still yet to be born!
Sometimes technologies, like our own kids, and the weather, surprise us.
I think we’ve been pleased by how quickly mobile has spread. The statistics suggest that most of the world has access to a mobile signal, the majority own phones, and about half are online. Despite these milestones, of course, challenges remain. There are gaps in access and use by women, by language, and by region. Some regions and sectors are racing ahead while others are falling further behind in the margins of poverty. We wouldn’t be here if technology had “solved” development challenges, and most of us will go back to our projects and our studies an our companies with a heap of work — a decade of work at least — still to do.
But I do have a question about those signs. In 2008 did you expect the confluence of mobile and social media, particularly Facebook, to be as central to digital development as it has become? Facebook now has over two billion users, more than a billion of whom log on each day. In some places it even has users who don’t know they are on the internet when they use Facebook.
Our work mapping the digital lives of people in Ghana, Kenya, and Uganda in 2015 uncovered mobile-first digital lives that were concentrated in platform use. Millions now get by online without an email address, relying on low-cost WhatsApp for everything from media sharing and church groups, to informal enterprise and civil disobedience.
This year as part of some work on digital ID supported by the World Bank, we met a young man, Cedric, in Cote d’Ivoire. Cedric was registered for a national ID, mostly so that he could register a SIM card and thus get to (you guessed it!) Facebook. Cedric’s motivations for obtaining a national ID would have been hard to predict in 2008. We weren’t yet attuned to how the intersection of mobiles and internet platforms would come to represent the new technology stack.
So how do we look forward, from the mobile/social/halfway there world we’re in now, to the challenges that we — as development practitioners and researchers — will face over the next decade?
When setting their trap for me, the conference organizers offered this framing question:
With entrepreneurial drive, the right technological solutions and a supportive enabling environment, what could digital development look like in 2028 for base-of-the-pyramid populations?
I can’t answer that question. To be more precise, I can’t offer a singular answer to that question. Instead, I’ll offer three. Depending on whether you are an optimist or a pessimist, a technologist or a sociologist, you can choose whichever answer that best suits your frame of mind.
Kidding aside, my goal is to help us to think about how these different perspectives are in productive tension with each other, how they are each a part of the digital development story, and how digital development will remain a fascinating, important, and complex domain in which to work.
The next 10 years will be a great time for digital development projects
I saw a tweet from Mobile World Congress the other day saying that the Ebola crisis marked a sea-change in the World Food Programme’s emergency response because for the first time the only way to reach some vulnerable populations was via mobile phones.
It’s just an anecdote, and mobile was far from a silver bullet for the crisis, but this is the type of story you want to hear if you are among those of us who are excited about using technology X to address development challenge Y. In this case, the fact that there was increasingly reliable cell coverage in many afflicted areas allowed for a more coordinated response. It also allowed emergency responders and researchers to gather more data about the spread of the disease.
When it comes to meeting any sustainable development goal, digital technologies play roles like these. Thus it’s both appropriate and encouraging that the communities of practice around ICTs for agriculture, education, livelihoods, water and sanitation, etc. are thriving.
And there’s more on the near horizon! Low Earth Orbit satellites may improve connectivity. Advances in user interfaces and natural language processing will allow people with a variety of literacies and limited computer skills to interact with institutions, with each other, and with their communities in new ways. And sensors and the Internet of things have the potential to accelerate agricultural yields, improve health and medical outcomes, transform logistics, and support smarter cities.
It’s a great time to do “X4Y.” We can now capitalize on all the infrastructure we’ve built over the last 10 years and layer on permutations of new technologies to address huge challenges in development, like the SDGs. Why wouldn’t we want to stay involved? So my first response to the organizer’s question is this: Dear colleagues, please don’t stop what you’re doing! Let’s continue to learn from each other, follow the principles for digital development, balance a focus on technology with its enabling environment, and take full advantage of technology to make people’s lives happier, safer, and more prosperous.
Access is necessary but insufficient for an inclusive digital economy — in the next 10 years we must push for universal effective use
In some ways, however, the task ahead may actually be more difficult. We can’t do X4Y without universal access and effective use of ICTs. And yet the last billion (or two) may be the most difficult to include.
One reason for this is the distinction between access and effective use. In this vein, I’m building on Michael Gurstein and several other scholars who have been arguing for many years that people benefit most from the Internet when they can use it in the way they want, to pursue the goals they want.
But we are only close to achieving near-universal access because most of the world’s population, particularly the most vulnerable, access mobile networks using pay-as-you-go minutes and data. This creates what in my book I call a metered mindset, in which people don’t surf and browse, but rather sip and dip. And that’s why our data from Kenya suggests that about 25% of Android users there take advantage of airplane mode — likely far more than have actually been on an airplane! A phone in airplane mode is not being used effectively.
I’m thrilled about efforts like the Alliance for Affordable Internet to promote targets of 1 for 2, one GB of data for 2% of average household income. But you see the tension in this math. Those with the least capacity to pay are those that the development community most needs to reach.
Other efforts to mitigate or subsidize the “metered mindset” have been rather ham-handed. Some might recall the fights around Internet.org, Free Basics, and zero-rating in India, as well as the preferential treatment of some Internet content over others. India prohibits zero-rating of this kind, but Free Basics is live in several other countries, and the disaggregation of internet sites into bundles is alarmingly widespread. So while we wait for the Low Earth Orbit satellites or the drones or weather balloons to change this, we’re stuck with two very different types of Internet users: those who sweat the next click, and those who don’t. For this reason, we need to be open to creative approaches like universal free 2g data, to promote universal access and unmetered, effective use among populations for whom there is not now and probably never will be a clear market rationale to serve.
And it’s not just about cost. The last billion are particularly hard to bring online not just because they are rural (and thus expensive to serve), but because they are disproportionately poor, disproportionately low-literacy, and disproportionately they do not speak one of the top 20 global languages. In some cases they are female members of families wherein the household power structures precludes them from the opportunities to use ICTs that men in the same families have, so gender divides persist.
And, digital literacy remains elusive — training at a massive scale is something we’re going to have to get better at.
And what internet will the last billion use? The centrality of business models in supporting or discouraging universal effective use is hard to overstate. Just as the ISPs, particularly mobile ones, landed on pay-as-you-go pricing as the best of a series of less-than-perfect solutions, the user facing Internet landed on advertising as the best of a set of not so great ways to support content-creation and provide more sophisticated services.
My colleague Bryan Pon at Caribou Digital did a study looking at the global market for advertising, and by placing the same ad in many markets he documented that ad rates in Africa are sometimes 1/20 of what they are in the US. This discrepancy makes it difficult for local media to go digital and support themselves via advertising, and it calls into question whether ad-supported content models are sustainable in the developing world.
Another Caribou colleague, Chris Locke, has suggested that at least in Africa, the advertising-led search and social media models that have worked well for middle-class populations may be stalling and that transaction led models, such as Go-Jek in Indonesia, which started as a motorcycle transport company rather than an advertising firm, may be better suited for the Internet of the future in the developing world.
I don’t have easy solutions for these problems. Instead, my second answer to the organizer’s framing question is that we’re still searching for the right ways to go beyond access, to support low-resource folks to use the internet and other technologies just as effectively (and frequently) as other people do. We have pretty good business models for delivering digital services and content to the emerging global middle classes today, but we have yet to find really good business models for delivering similar levels of content and services to the very poor.
We’re building the network society, not just connecting people to it
In my third answer to the question, I should find a path between the optimism of the first answer and pessimism of the second. So I’ll raise what I think will be one of the dominant debates in our field in the next 10 years.
As we look ahead to the next decade, the digital development community needs to be cognizant of its roles in creating and shaping the digital world writ large. From digital ID to digital financial services, gig economy work patterns to social credit scores, the world we’re helping to build and shape is re-configuring local and global, rich and poor. This transformation goes by many names, including the network society, the information society, or the digital age. Regardless, this digital world is being built as we speak, shaped by decisions made in Nairobi, Bengaluru, and Shanghai; and particularly Menlo Park, Mountain View and Seattle, with or without the input of the “digital development” community.
The reason I called attention to platforms is that I can’t say with confidence that we all believe that our own circumstances are better because of the concentration of these digital entities. And, if we don’t think that we are necessarily better off, we need to be mindful about promoting digital policies and programs that depend on the power of these platforms in the developing world.
Platform power has real implications for privacy, for national sovereignty, for local media systems, and for strong civil societies. It also has implications for the ability of people to work with dignity and confidence and earn enough money to support their families while competing with machines, or algorithms, or people halfway around the world for 20 minutes at a time. Even app developers face difficulties with platforms like app stores; a recent Caribou Digital analysis illustrates how app store rankings help privilege some providers over others, entrenching US app producers instead of surfacing local producers. Platforms aren’t all bad, but they are a major driver of the labor and civil society disruptions outlined in the 2016 world development report.
Disruption and technology go hand in hand, but, as the digital transformation accelerates, it is our responsibility not just to talk about ICT in employment but to talk about ICTs and what kind of employment. Not just to talk about access or a digital divide but about effective use. Not just to talk about airtime top-ups and P2P transfers but to talk about meaningful financial inclusion. Not just to talk about ICTs and participation but about what kinds of discourses and what kinds of public information systems we want to build.
Twenty-five years ago, Sir Tim Berners-Lee described the World Wide Web as “a pool of knowledge that is as easy to update as it is to read,” a place where everybody has an equal opportunity to be both the producer and consumer of information. Let’s remain aware of this vision. Our deployments, even if they make sense locally, may be reflective of a broader shift in the nature of the Internet — a shift which we may not be entirely happy with. So if you have the choice to just run something on WhatsApp group or on Facebook, because that’s where the users are, you might want to think about keeping that standalone web page open.
The other critical issues lay ahead in machine learning and artificial intelligence. It is crucial that we develop AI in a way that works for the whole world, not the global north. One example is to ensure that language processing works as well in smaller languages as does in Spanish or English. Another might be the development of ethical, privacy-respecting predictive models for citizen behavior and credit scoring. These aren’t mere technical questions or funding gaps — that 40% of the world’s poor is likely to be left behind by the advances in NLP reflects a set of increasingly interconnected forces. Global firms at the heart of the Internet have to make choices about deploying resources and investments into very different environments. Will those choices promote widespread inclusion and effective use? What power discrepancies do those choices create?
You can tell by now that I’m advocating for an awareness of how the systems we’re building and the technological stacks we’re are utilizing, increasingly, encode the terms of participation in a global age. To answer the question a third time, I’d say that we have a tremendously important role to play in expanding the agenda of digital development beyond one of ICT4D’s technical solutions to narrow problems or the digital divide approach to access gaps. The field has not been very good at this, at least not yet. In our workshops and pilot papers and proposals and application of digital principles, we must better attune ourselves to the world we build when we code.
“Digital development” is a new-ish term. It’s refreshing not to be at an “ICT4D” event and very encouraging to hear some influential institutions using “digital development”, and even the more holistic “doing development in a digital age” this year.
Yet as I close, you’ll note that it took three answers to build this vision
a) an invigorated “X4D” supercharged by new technologies
b) a pragmatic not-just-market approach to “universal effective use” (not just access), and
c) an increased awareness that we’re building a global network society, not just connecting people to it.
Actually I didn’t want you to choose. The trick for this vision to work for the digital development community is for it to hold onto all three elements at the same time, even if individually, we each may have our favorite bucket.
Thus the difference between 2008, 2018, and 2028 will lie as much inthe quality of the interplay between these elements as it will in any new technological fix. Together these elements point to a diverse community of practice more in alignment with the digital age and better positioned to shape the direction of global development.
We’ve got a lot of work to do, so let’s keep at it!
Thanks so much for your time and attention today and for the chance to think through some of these issues with you.