Access for Everyone: A Model for Free Online Learning, with Duolingo's Luis von Ahn | Big Think
The problem with language education, there's about 1.2 billion people in the world learning a foreign language. It's one of the most common things that people learn in the world, everywhere in the world except maybe for the U.S. It's not that common in the U.S., but everywhere else it's about 1.2 billion people learning a foreign language.
Now, if you look at it more deeply, it turns out about 800 million of them satisfy three properties. The first one is they're learning English. The second one is that they are doing so in order to get a better job or a job at all, and the third one is that they are of low socioeconomic condition. So basically, most people learning a foreign language are poor people learning English to make more money or to make some money.
Now, the kind of ironic thing is that usually the way there are to learn languages, and particularly to learn English, costs a lot of money. So for example, in the U.S. there's Rosetta Stone, which is $500-$1000. In Latin America there's a program called Open English, which is about $1000. So it's this ironic thing that most of the people that need to learn a language are poor people that are doing it so that they can get money, but it requires quite a bit of money to do so.
Which is why with Duolingo we decided to make a completely free way to learn a language. And that's the whole premise of Duolingo. When we started, we thought we have to make a way to learn a language, but it has to be 100 percent free. When we started Duolingo, it was not just me. It was me and my co-founder whose name is Severin Hacker, who is very funny because his last name is Hacker.
When we started, we knew we wanted to do a free way to learn languages. This is what we wanted to do. It's easy to say that it's free; the problem is when something is free you got to find a way to finance it and to make it sustainable. So the question really became, how do we teach languages for free but such that we can actually finance the whole thing?
The solution to this came from many of my previous projects that had this very similar idea. And it's an idea that can be traced back to an idea that I had when I was a kid. It was a terrible idea, but at the time I thought it was an amazing idea. And it was that I wanted to have a gym where it was free to go to the gym. It's a free gym, but all the exercise equipment was connected to the power grid, and people, when they went there as they exercised, would generate electricity that the gym would sell to the power grid. So that's why it was free.
We wouldn't charge people, but we would make money by selling electricity to the electric company. It turns out this is a bad idea because it turns out humans are actually not very good at making electricity. But I thought it was a good idea at the time. Also, there's another reason why it's a bad idea. It turns out with gym economics actually, most of the money is made from people not showing up, whereas in this case, we really needed people to show up because we needed to generate the electricity.
But it is not a good idea, but it's a very similar idea to what we ended up doing with Duolingo and what I have used in my previous projects, where the idea is, can we offer a service for free, but as we're offering it for free, can we extract some value out of people doing the thing anyways?
So in the case of the gym, it was extracting electricity. The question is, what value can we extract out of people learning a language? And it turns out what you can do is you can get people who are learning a language to help with language translation. So what we do with Duolingo, the way we finance Duolingo is that whenever we teach somebody a lesson, so we may teach them about food words in a given language, at the end, once we've taught them about it, we say, "Hey, if you want to practice what you just learned with something from the real world, here's this document that has never been translated before that is in the language that you're learning. Can you help us translate it to your native language?"
And then we sell that translation. So for example, CNN is one of our clients, where the idea is that CNN writes everything in English, they send it to us, then people who are learning English on Duolingo, in order to practice their English, they get a CNN story in English, and they have to translate it into their native language.
Now we get multiple people to translate the same story, and they vote on each other's translations. And at the end, we come up with one best translation for the whole article, and then we send it back to CNN, and CNN pays us for having translated their article. So that's the idea; people as they're learning a language, they're helping us do the translation, and thus they're practicing.
It's completely optional. You can learn everything there is to learn on Duolingo without ever practicing this way. But about half the people who get presented an opportunity to translate say they would like to translate it. So it ends up working out. Ever since we launched Duolingo, it has grown a lot. We now have 42 million users, from zero to 42 million users in two years.
We are the most popular way to learn languages in the world. There, in fact, in the United States, there are more people learning a language on Duolingo than in the entire U.S. public school system. So we have a lot of people learning a language. They're also not paying.
And also in the developing world, there are millions of people who before were just not able to learn a language; they didn't have access to it, whereas now with Duolingo, they do. So we've already changed the industry. If you look at, for example, the stock price of Rosetta Stone, it's really not good. But I think we're only getting started.
I think we're nowhere near as good as I want to be. I think we should be able to teach you a language three, four times more effectively than we do in terms of the time it takes to learn it. And so we're going to be working on that, on really making it so that we are your one-on-one tutor, but it's a computer one-on-one tutor, and I think we'll be able to do it.