Bill Gates on Mobile Banking, Connecting the World and AI
A big bet on improving the lives of the poor—and, as always, an abiding belief in software
Fifteen years ago, Bill and Melinda Gates began a foundation that would spend billions to improve the lot of the world’s poorest. They have made a considerable impact, most notably saving more than an estimated five million lives (and probably plenty more since that 2012 approximation.) With this year’s annual letter to the public, Bill Gates looks forward to the next 15 years and makes a big bet that the lives of the poor will improve even more dramatically, citing four breakthroughs in particular: lower child mortality, improved agriculture in Africa, transformative educational software, and life-changing mobile banking that will radically improve the lives of the poor. The latter initiative is important given that 2.5 billion adults do not have access to financial institutions; among the extremely poor, only 20% do, making even the simplest transactions arduous if not impossible. On the eve of the release of the Gates Letter, Backchannel spoke to the philanthropist about banking and other issues, including the former Microsoft CEO’s view of current concerns about the threat of artificial intelligence.
[Steven Levy] In this year’s Gates Letter, you seem to be focusing more on the ubiquity of connected technology than in previous years. Is that a fair characterization?
[Bill Gates] Yes and no. With the breakthrough in education, although it’s enabled by getting the networks out in poor countries and getting the phones and tablets out, the big thing I’m talking about is the quality of online software: personalized learning, great lectures, forums that people can engage in, and projecting out where we’ve come— from just putting a video camera in front of a lecturer in the beginning, to where we are today with Khan Academy and the other good sites that are personalized and interactive.
I also have the benefit in this letter—which has not been true in other ones—of taking a pretty long time frame. Fifteen years is a long time. On banking, for instance, you could say I’m assuming the network and the mobile phone [will be ubiquitous], otherwise those things can’t happen. But what the foundation does [in the meantime] is getting the regulations right, getting the innovations of financial savings and loans for farmers to get seeds and fertilizer, [for parents] to pay school fees. Those wouldn’t be possible without the hardware base. But the magic new element is the software aiming at those particular needs. That’s where our foundation has been the biggest funder.
In the past you’ve actually mocked some global connectivity projects, like Google’s Loon Project and Mark Zuckerberg’s Internet.org. Are you taking a different view of those efforts now?
The key is not Internet connectivity. There’s plenty of places where you have Internet connectivity. Urban Africa has Internet connectivity. So, yes, Internet connection is very enabling for some of the things we get involved in. It doesn’t in itself stop childhood mortality, or help farmers know what to plant. But the fact that all the good knowledge can get out in this digital world is very, very exciting. If it wasn’t for the digital revolution, we could not be so ambitious. But if you just stood back and watched those things take place, the children would still be dying of malaria and the poor would still have no banking.
One of the four big initiatives you talk about in the Notes is mobile banking. The Bill and Melinda Gates Foundation has funded projects like bKash in Bangladesh and M-Pesa in Kenya that allow people who have never had checking accounts to do banking with cheap mobiles phones. Does your involvement go beyond funding and into product design?
The actual ideal is to essentially have a public switch that lets you move funds at low fees between various banks. That kind of rich interchange capability actually still doesn’t exist. So even though M-Pesa has gone a long way it’s not the ideal platform in terms of ease of entry and ease of competition. Also as it gets to be a big system, the kind of robustness that you need to have is going be subject to a lot of attacks. So yes, there’s a lot of deeper technology that has to be brought to bear here to get the ideal system. And M-Pesa proves a lot but it’s not like all we have to do is just take exactly that and put it in all these other countries. We built up [an organization with] some good people. They’ve done a fantastic job of bringing in outside experts, like talking to the PayPal guys— Max Levchin is giving us advice. In terms of getting advisory boards—people very deep in the technological piece—we’ve done as good a job in the financial services area as any area that we’re in.
Are you excited about the potential of Bitcoin in systems like these as a way to keep fees low and have the system work robustly in the global sense?
There’s a lot that Bitcoin or Ripple and variants can do to make moving money between countries easier and getting fees down pretty dramatically. But Bitcoin won’t be the dominant system. When you talk about a domestic economy, [you must have] the idea of attributed transactions, where if you sent it to the wrong person you can actually get the transaction reversed. [And a traditional system] doesn’t have this huge fluctuation where the value of your account is going up and down by a factor of two. We need things that draw on the revolution of Bitcoin, but Bitcoin alone is not good enough.
The Gates Foundation has a video of transactions in the undeveloped world five years from now that seems much more advanced than what’s in the United States now. Will poorer African and Asian countries actually leapfrog us by widely adopting these mobile transactions first?
In the rich world the switch to pure digital, which should have a lower fee structure, hasn’t happened. The average transaction size is high, partly because of the complicated way that banks have bundled products together. Our challenge is to make very small transactions, well under a dollar, be able to happen in huge volume and yet bear less than a two percent fee. So we’re a little bit more demanding. It does mean there’s a chance in some of these [other] countries—particularly if we get this thing to critical mass in India. Our top two priorities, although we’re working in dozens of countries, the two that we’re putting the most resources into right now are India and Nigeria. If we got to critical mass in those two countries, that would be really dramatic.
Your foundation has funded systems with low fees, but will they be self sustaining?
Yeah. We’ve done the bootstrap. As you get the volume of transactions up, even a one percent fee can be very attractive in a high-volume, at-critical-mass system.
There is a trend in connectivity called zero rating where certain essential digital services, like Wikipedia, are not charged against the data rates that a user must pay. Do you think banking should have that status? Are you a supporter of zero rating?
Zero rating is kind of the old closed garden concept. Maybe that’s an okay bootstrap, but for an entrepreneur who has got a better way of doing agricultural loans or somebody who’s got different courseware, the let-the-best-product-win approach is general Internet capability. So unless you’re really saving somebody cost in a big way, then it’s just a weird form of price discrimination. My general view would be, “Hey, let’s give them general Internet,” and maybe the pricing should be tiered based on income levels.
Let me ask an unrelated question about the raging debate over whether artificial intelligence poses a threat to society, or even the survival of humanity. Where do you stand?
I think it’s definitely important to worry about. There are two AI threats that are worth distinguishing. One is that AI does enough labor substitution fast enough to change work policies, or [affect] the creation of new jobs that humans are uniquely adapted to — the jobs that give you a sense of purpose and worth. We haven’t run into that yet. I don’t think it’s a dramatic problem in the next ten years but if you take the next 20 to 30 it could be. Then there’s the longer-term problem of so-called strong AI, where it controls resources, so its goals are somehow conflicting with the goals of human systems. Both of those things are very worthy of study and time. I am certainly not in the camp that believes we ought to stop things or slow things down because of that. But you can definitely put me more in the Elon Musk, Bill Joy camp than, let’s say, the Google camp on that one.