Posts Tagged ‘Boston Review’

Ideas Matter: Can Technology Solve Global Poverty?

December 3, 2010

On December 2, 2010, there was a panel discussion titled “Can Technology Solve Global Poverty?” in Cambridge, MA, hosted jointly by the Boston Review and the MIT Political Science Department. (The Jester appreciates the platform provided by the Boston Review — thanks to editors-in-chief Josh Cohen and Deb Chasman!) The panelists were Kentaro Toyama, Nicholas Negroponte, Rachel Glennerster, and José Gómez-Márquez, and moderation was handled expertly by Archon Fung. Brief bios of the panelists are all available here:; a video of the event should appear there soon, as well.

There was some drama on stage, but the panelists’ views were nothing new for people who follow this space. So, just a quick summary…

– Toyama: Technology magnifies human and institutional intent and capacity. In international development, technology is rarely a solution by itself. (Geez, this guy is a human broken record!)

– Negroponte: Laptops transforms education for children. Anyone who can’t see this needs a therapist!

– Glennerster: Look for whatever solutions work in international development, technology or otherwise. Then do randomized control trials (RCTs) of them to verify effectiveness.

– Gómez-Márquez: We need to design technologies and systems so that they will work in a hostile environment.

Overall, the Jester couldn’t have agreed more with Toyama, but there were some things that Toyama didn’t do well. For example, he didn’t clarify that he was talking mostly of ICT up front (although the Jester increasingly believes the amplification thesis applies beyond ICT and beyond international development). He also came  off as anti-technology, or anti-ICT4D, which is not quite true. He’s just pro-foundational-investments-in-human-capacity-that-rarely-require-much-technology.

Glennerster and Gómez-Márquez both held extremely reasonable positions about technology, namely that sometimes they can be helpful. The only problem with extremely reasonable positions is that while they are invariably true, they provide no additional insight.

Toyama’s core thesis primarily leads to arguments against (1) the indiscriminate spreading of technology without a full understanding of impact, (2) any hype around technologies potential that ignores the necessary human or institutional requirements for success, and (3) ignoring of opportunity costs when cheaper solutions abound. There is nothing wrong with cost-effective technology being used in the right way to amplify existing positive intent and capacity, as the Jester recommends for ICT4D-ers.

Of these, Point (3) on opportunity costs was discussed on the panel. (Everyone except for Negroponte appeared to agree with (1) and (2).) The Jester fully agrees with Toyama’s point about opportunity costs. When presented with multiple ways to solve similar problems, which should you choose? One slogan often goes, “It’s not either/or, do all of them!” This is an approach that might appeal to the United Nations, where participation and consensus is the goal, but the reality is that funds allocated to international development are always limited. If you “only” have a budget of $100 million for a million students, you can’t both buy them all computers and do meaningful teacher training. You have to choose how to allocate the budget.

That choice often, though not always, comes down to a simple question — Which intervention provides more bang for the buck? Unfortunately, ICT rarely comes in on the cheaper side, particularly in low-labor-cost environments (note to self: Jester, hurry up and get to Myth 9: “Automated is always cheaper and better”). Negroponte seemed excited to reveal that OLPC only costs a dollar a week per child, but as Toyama and Glennerster both responded, there are interventions that cost 100 times less, with known and significant educational benefits, and even a dollar a week is too much for countries that barely spend that much total on education per child. (Actually, the Jester even doubts Negroponte’s dollar a week number — note to self: hurry up and get to Myth 8: “Hardware and software are a one-time cost”!)

Glennerster and Gómez-Márquez were both very persuasive in the importance of designing interventions well, so that they work even in hostile environments. Toyama had difficulty responding to this point, although he should have done better considering that he used to make similar remarks himself. Gómez-Márquez, in particular, made an intriguing comment that you could design things so that they work even under adversarial conditions. This sounds fantastic in theory, but in reality, someone somewhere in the system must have the intent to solve the the problem, for any technology to work (this is similar to what computer security people say about computer security). It might be a minister, an NGO leader, a local entrepreneur, a group of mothers, or some combination, but every technology requires positive human capacity behind it to activate.  (In a post-panel conversation, Gómez-Márquez acknowledged this point.) Gómez-Márquez talked about identifying what might be called “champions” in a given environment, and then providing them with the right tools. Toyama would undoubtedly agree with this approach (the Jester agrees, too) — it means that the technology is amplifying the champions’ intent and capacity. ICT4D projects whose stated intent is to identify and amplify champions make perfect sense.

Glennerster brought up the topic of vaccines, which are a theoretical pebble in the Jester’s pointy shoes. Such technologies do pose a partial counterexample to the theory of technology as amplifier, because at the least, they don’t amplify negative intent (unless, the Jester supposes, some of them can turn to poisons when not used as indicated). But, even vaccines are subject to bad institutional capacity, and they are certainly not immune (ha ha — the Jester should be punished for every pun he sheds) to the amplification thesis. In the end, vaccines are regularly and routinely distributed unequally, which is exactly why yellow fever still exists in the developing world, despite the technological existence of reliable vaccines. Disproportionately, it’s poorer countries that continue to have these problems and it’s usually because the vaccine supply chain is not in order.

Of course, none of this says that we shouldn’t develop the technology — the Jester so far hasn’t suggested that PCs or mobile phones should be uninvented. (Not yet, anyway!)

Finally, the Jester came away with new appreciation for Negroponte’s persuasiveness, if not his logic. Although Negroponte lost his cool at times (and perhaps did more to hurt his own cause than to advance it), he does genuinely appear to believe 100% in the power of laptops alone to radically transform children’s education for the better. And possibly as a result, his tone, if not his rational argument, is incredibly seductive. The Jester found himself nodding along hypnotized, while Negroponte compared laptops to vaccines and suggested that nothing that required experimental evaluation was worth doing.

These last points will be addressed in the Jester’s next post. The Jester passed on commenting on Negroponte in a previous post, but it seems necessary.  Negroponte’s salesmanship requires repeat doses of a rational antidote!

Baseball Analogies

November 17, 2010

In his essay in the Boston Review’s recent forum on “Can Technology End Poverty?“, Archon Fung writes, “What matters is hits and runs rather than the batting average.” The Jester prefers juggling balls over pitching and batting, but let’s see where the analogy takes us.

In Fung’s use of the analogy, the idea is that development practitioners using technology are baseball players and what we want is to increase the likelihood of scoring runs. He says, “Toyama’s concern, then, is whether the batting average can be increased.” The Jester has it on good authority that this is actually not quite what Toyama had in mind (though it is certainly a secondary issue). In fact, there are several problems with this formulation.

First, it completely buys into a model where development practitioners are engaging in all the action and that poor people are just passive recipients. In fact, in Fung’s use of the analogy, the beneficiaries of development don’t even appear on the field as human beings. (They are certainly not the opponent team.) Any positive impact in development is measured in runs batted in.

Second, the analogy assumes a flavor of technological determinism: Some hits are good, some are bad. (“As long as someone invents an iPhone for every few Apple Newtons or Microsoft KINs, technology will continue to improve human welfare.”) There is no subtle consideration for the manifold and complex role of technology. In effect, Fung is saying that while not all technologies are necessarily good, some are necessarily good. This is different from what Toyama says, which is that technology is effectively neutral, but whether it’s used well or not is up to people. (Toyama’s article also reduces technology to a simple amplification, but if he had had more space, he would probably have gone into more detail about how complex human intent and capacity interact with technology to cause complex outcomes. Probably.)

Fung then continues that because we don’t know which technologies will have good or bad impact to begin with, we need to try lots of them. Of course, this again assumes a kind of technological determinism. He then states that a technology’s efficacy is the key factor to focus on. “What are the problems that telecenters or plastic laptops are supposed to solve, and how are they supposed to solve them?” The Jester finds this a convenient place to say, social problems aren’t solved by technology. They’re solved by people. Technology may amplify people’s capacity to solve a problem, assuming they are already inclined and able to solve the problem to begin with.

All of this technological determinism leads to Fung’s recommendation that we should focus on technological affirmative action: Design the technology so that it is progressive in nature. (A variation of this argument was raised by Pedro Ferreira in a response to an earlier Jester post.) This conception, however, is flawed because there is no such thing as a technology that is progressive in itself. Above, there was the claim that “technology is neutral.” This isn’t strictly true, as many scholars of technology are keen to point out; technology is not “value free.” For example, it’s difficult to use an effective vaccine for anything other than a positive purpose. And, it’s hard to imagine constructive uses of thumbscrews. True enough. But, even these uses ultimately only magnify what human intent and capacity are behind them. Their lopsided value is due to lopsided magnification, not from inherent intent built into the technology. The WHO estimates that 1.4 million children under five die of vaccine-preventable diseases. The technology exists; intent and capacity are deficient. Conversely, thumbscrews have lost their popularity, as global intent to torture appears to be declining (frequent infractions by various parties notwithstanding). Again, the technology exists; bad intent has lessened.

To summarize, even a progressively valued technology only results in progressive outcomes if the human intent behind it is progressive. What does this mean for ICT4D? First, an indiscriminate spreading of technology rarely accomplishes a progressive end in those places where you’d like to see progressive goals reached, because exactly in those places, progressive intent is lacking (if it weren’t lacking, you wouldn’t have great inequality to begin with; this point isn’t as black and white as stated; there is a grey gradient, but you get the idea). Second, the only way to achieve a progressive end with technology is for progressively minded people to put technology to progressive use. This is, in fact, what many ICT4D projects attempt, and to that extent, they make sense. The Jester simply wants to nudge more of that progressive intent to be free of the constraint that technology must be used to solve the underlying problem, as well as to focus more on building human capability. What doesn’t happen is a technology by itself, however great and positive-leaning the design, solving a social problem simply by large-scale dissemination.

The Jester argues for a completely different use of the baseball analogy: The question of ICT4D is about whether a better bat will result in the worst players hitting better, thereby closing the the gap between good and bad players. As to closing the gap, the Jester believes, basically not at all. As to improving the hitting ability of bad players, the Jester says, possibly, depending on how bad. If the players are malnourished, exhausted from a day’s labor, blind in one eye, and have never played baseball, it’s very unlikely; they’ll strike out no matter how good the bat. Between players who don’t have those severe problems, a better bat will help everyone, but it will help better players more. Of course, you could progressively give the worst players the best bats. That will help, but Ichiro will still outbat the Jester. (And, you can’t really keep the best bats out of the hands of the best hitters, anyway.)

Ultimately, all the focus on the bat seems a little weird. What you really want is to feed the malnourished players, and give intensive batting lessons to all the bad players. Not everyone will be Babe Ruth, but everyone will have a decent shot at the plate. The best affirmative action isn’t to waive considerations of merit as a gatekeeper, but to intensively nurture capacity in those who lack it well before the gate, so that everyone has a chance to be accepted.

What We’d Like to Hear from OLPC

November 12, 2010

For the next few posts, the Jester will address both public and private discussions ensuing the Boston Review’s forum essays on “Can Technology End Poverty?“.

Nicholas Negroponte’s response went up first. At first, the Jester was going to do a point-by-point repartee, but after the second paragraph, he realized it would take too long and sound too much like his Mad Lib post. So, instead, the Jester uses his ventriloquist skills and puts words into Negroponte’s mouth: What we’d like to hear from OLPC…

– – – – – – – – – –

Laptops Work… but Is That Education?
Nicholas Jesterponte

It’s great that such an important topic is being discussed with this lead article (and by such a handsome, intelligent guy to boot). It put to words some unease I had been having about things recently.

When I started One Laptop per Child (OLPC) in 2004, I said that owning a connected laptop would help eliminate poverty through education, especially for the 70 million children who have no access whatsoever to schools. It’s hard for me to let go of this belief.

But, what I have learned after several years of feedback from governments, educators, and critics, as well as our own experiences in several countries, is that the operative word here is “help.” You actually need a lot more than laptops to change the face of education in the world. You need dedicated teachers, good administrators, involved parents, and committed politicians. Any technology needs to be integrated into the curriculum, and you need to train teachers in their use. On an ongoing basis. Of course, you need great students, too, but I find all young children are remarkable in their capacity to learn.

Kentaro Toyama is coming from a place that I refused to acknowledge at first. In fact, the acronym, OLPC, shows where we were coming from back then. Any name of the form OXPY ought to ring alarm bells. Is the intent simply to provide every Y with technology X, or do we really want to see meaningful development outcomes? The name implies that our mission was mass disseminate technology, not real education. I’ve kept the name, but more because it’s a known brand. For me, it serves as a reminder of our early missteps.

For example, one of my initial hopes was to get India hooked on OLPC. But although I was traveling to India five to six times in those days, they didn’t go for the 1 million laptop deal I presented to them. At first, I couldn’t believe it! Laptops are a wonderful way for all children to learn learning through independent interaction and exploration. How could they not get that? It seemed so obvious to me. Maine had a great laptop program that did really well, and I had so many brilliant friends at MIT who grew up tinkering with computers.

Surprise! It turns out that India isn’t Maine and that MIT professors are a little atypical. Indian government schools are an eye-opener. I’ve seen schools where class meets under a tree and students do arithmetic in the dirt. I’ve seen schools where most of the teachers don’t show up. I’ve seen schools that have no toilets for the children, and so then, the girls don’t show up. And that’s even though you can build a decent toilet for $300 there. When I saw these things, I started to wonder a bit about the appropriateness of a $188 laptop.

In fact, I later learned that Indian public spending on primary education is between $70-$200 per student per year, and that most of it goes to teacher salaries. How could we then ask that they then pay $188 per laptop? (And, the $188 doesn’t include incidental costs that are bound to come up, like those of maintenance and teacher training.) Even at $1 per week, OLPC is pretty expensive! As Toyama notes, many private schools in India only charge $1 a month.

In retrospect, I feel glad that India didn’t spend $188 million or more on our laptops. I’m not sure I would have wanted that on my conscience.

At this point, we have over 2 million laptops out there, but I’m glad to report that most are in countries that are wealthier than India on average. Peru’s per capita GDP is around $4500 (they have 300,000 laptops) and Uruguay’s around $10,000 (400,000 laptops). They also each spend more of their per capita GDP on each primary school student. I’m a little worried that Peru wasn’t quite ready for us, but Uruguay is likely to be OLPC’s star country.

Uruguay, first of all, starts with a great educational system. Literacy is high, at over 98%. The commitment to education is strong. As soon as they decided to try laptops for all of their students, they were able to get Internet connectivity to 98% of their schools! Phenomenal. (In contrast, less than 10% of Peru’s schools have managed connectivity.) They’ve invested in great teacher training and they have strategies to develop relevant content. The teacher training programs are sophisticated with support teachers, master teachers, and ongoing teacher meetings.

I’ll admit even in Uruguay, all is not rosy. Between 25% and 35% of the laptops are effectively out of service at any given time, and maintenance is hard to do at a national scale. Also, we really need to do an evaluation there, to understand exactly what kind of impact OLPC has had beyond the anecdotal stories that my team loves to tell me about. And, that’s not even beginning to address questions of cost-effectiveness.

So, what have I learned from all of this? If you’re a country, you need a good educational system in place and a budget that isn’t overwhelmed by the $188-per-student cost, before you should even think about OLPC. And, if you’ve got those things, you’ll need to think still harder. Remember that in addition to the first batch of hardware, you’ll need to invest in teacher training, connectivity, curriculum development, maintenance, replacement costs, and oh, did I mention teacher training? All of these things can add up. (Note to self: We should check back on Uruguay to see how much the entire program is actually costing them.) After all that, if you’re still interested, by all means, call me… I have some laptops to sell you, as well as some tips for best results. Among them: start small, and scale gradually. If it’s really worth it, you’ll get to a million eventually.

How do you eliminate poverty? The answer is simple: education. How do you provide education? The answer is less simple. I no longer believe that just a laptop per child will do it. Being a technologist, I can’t get it out of my head that somehow laptops ought to help. But over the years, I’ve come to appreciate that education is a deep, complex, social challenge, and that if OLPC is going to play a role, it will be in a support role, not as the main actor.

One thing I can say for sure: OLPC has been a great education… for me.