How many of you had to fill out a web form where you've been asked to read a distorted sequence of characters like this? How many of you found it really annoying?
OK, outstanding. So I invented that.
Or I was one of the people who did it. That thing is called a CAPTCHA. And it is there to make sure you, the entity filling out the form, are a human and not a computer program that was written to submit the form millions of times. The reason it works is because humans, at least non-visually-impaired humans, have no trouble reading these distorted characters, whereas programs can't do it as well yet. In the case of Ticketmaster, the reason you have to type these characters is to prevent scalpers from writing a program that can buy millions of tickets, two at a time.
CAPTCHAs are used all over the Internet. And since they're used so often, a lot of times the sequence of random characters shown to the user is not so fortunate. So this is an example from the Yahoo registration page. The random characters that happened to be shown to the user were W, A, I, T, which, of course, spell a word. But the best part is the message that the Yahoo help desk got about 20 minutes later.
[Help! I've been waiting for over 20 minutes and nothing happens.]
This person thought they needed to wait. This, of course, is not as bad as this poor person.
CAPTCHA Project is something that we did at Carnegie Melllon over 10 years ago, and it's been used everywhere. Let me now tell you about a project that we did a few years later, which is sort of the next evolution of CAPTCHA. This is a project that we call reCAPTCHA, which is something that we started here at Carnegie Mellon, then we turned it into a start-up company. And then about a year and a half ago, Google actually acquired this company.
Let me tell you what this project started. This project started from the following realization: It turns out that approximately 200 million CAPTCHAs are typed everyday by people around the world. When I first heard this, I was quite proud of myself. I thought, look at the impact my research has had. But then I started feeling bad. Here's the thing: each time you type a CAPTCHA, essentially, you waste 10 seconds of your time. And if you multiply that by 200 million, you get that humanity is wasting about 500,000 hours every day typing these annoying CAPTCHAs.
So then I started feeling bad.
And then I started thinking, of course, we can't just get rid of CAPTCHAs, because the security of the web depends on them. But then I started thinking, can we use this effort for something that is good for humanity? So see, here's the thing. While you're typing a CAPTCHA, during those 10 seconds, your brain is doing something amazing. Your brain is doing something that computers cannot yet do. So can we get you to do useful work for those 10 seconds? Is there some humongous problem that we cannot yet get computers to solve, yet we can split into tiny 10-second chunks such that each time somebody solves a CAPTCHA, they solve a little bit of this problem? And the answer to that is "yes," and this is what we're doing now.
Nowadays, while you're typing a CAPTCHA, not only are you authenticating yourself as a human, but in addition you're helping us to digitize books. Let me explain how this works. There's a lot of projects trying to digitize books. Google has one. The Internet Archive has one. Amazon, with the Kindle, is trying to digitize books. Basically, the way this works is you start with an old book. You've seen those things, right? Like a book?
So you start with a book and then you scan it.
Now, scanning a book is like taking a digital photograph of every page. It gives you an image for every page. This is an image with text for every page of the book. The next step in the process is that the computer needs to be able to decipher the words in this image. That's using a technology called OCR, for optical character recognition, which takes a picture of text and tries to figure out what text is in there. Now, the problem is that OCR is not perfect. Especially for older books where the ink has faded and the pages have turned yellow, OCR cannot recognize a lot of the words. For things that were written more than 50 years ago, the computer cannot recognize about 30 percent of the words. So now we're taking all of the words that the computer cannot recognize and we're getting people to read them for us while they're typing a CAPTCHA on the Internet.
So the next time you type a CAPTCHA, these words that you're typing are actually words from books that are being digitized that the computer could not recognize. The reason we have two words nowadays instead of one is because one of the words is a word that the system just got out of a book, it didn't know what it was and it's going to present it to you. But since it doesn't know the answer, it cannot grade it. So we give you another word, for which the system does know the answer. We don't tell you which one's which and we say, please type both. And if you type the correct word for the one for which the system knows the answer, it assumes you are human and it also gets some confidence that you typed the other word correctly. And if we repeat this process to 10 different people and they agree on what the new word is, then we get one more word digitized accurately.
So this is how the system works. And since we released it about three or four years ago, a lot of websites have started switching from the old CAPTCHA, where people wasted their time, to the new CAPTCHA where people are helping to digitize books. So every time you buy tickets on Ticketmaster, you help to digitize a book. Facebook: Every time you add a friend or poke somebody, you help to digitize a book. Twitter and about 350,000 other sites are all using reCAPTCHA. And the number of sites that are using reCAPTCHA is so high that the number of words we're digitizing per day is really large. It's about 100 million a day, which is the equivalent of about two and a half million books a year. And this is all being done one word at a time by just people typing CAPTCHAs on the Internet.
Now, of course, since we're doing so many words per day, funny things can happen. This is especially true because now we're giving people two randomly chosen English words next to each other. So funny things can happen. For example, we presented this word. It's the word "Christians"; there's nothing wrong with it. But if you present it along with another randomly chosen word, bad things can happen. So we get this.
But it's even worse, because the website where we showed this actually happened to be called The Embassy of the Kingdom of God.
Here's another really bad one. JohnEdwards.com
So we keep on insulting people left and right everyday. Of course, we're not just insulting people. Here's the thing. Since we're presenting two randomly chosen words, interesting things can happen. So this actually has given rise to a really big Internet meme that tens of thousands of people have participated in, which is called CAPTCHA art. I'm sure some of you have heard about it. Here's how it works. Imagine you're using the Internet and you see a CAPTCHA that you think is somewhat peculiar, like this CAPTCHA.
What you're supposed to do is you take a screenshot of it. Then of course, you fill out the CAPTCHA because you help us digitize a book. But first you take a screenshot and then you draw something that is related to it.
That's how it works.
There are tens of thousands of these. Some of them are very cute.
Some of them are funnier.
And some of them, like paleontological shvisle ...
they contain Snoop Dogg.
OK, so this is my favorite number of reCAPTCHA. So this is the favorite thing that I like about this whole project. This is the number of distinct people that have helped us digitize at least one word out of a book through reCAPTCHA: 750 million, a little over 10 percent of the world's population, has helped us digitize human knowledge. And it is numbers like these that motivate my research agenda. So the question that motivates my research is the following: If you look at humanity's large-scale achievements, these really big things that humanity has gotten together and done historically — like, for example, building the pyramids of Egypt or the Panama Canal or putting a man on the Moon — there is a curious fact about them, and it is that they were all done with about the same number of people. It's weird; they were all done with about 100,000 people. And the reason for that is because, before the Internet, coordinating more than 100,000 people, let alone paying them, was essentially impossible. But now with the Internet, I've just shown you a project where we've gotten 750 million people to help us digitize human knowledge. So the question that motivates my research is, if we can put a man on the Moon with 100,000, what can we do with 100 million?
So based on this question, we've had a lot of different projects that we've been working on. Let me tell you about one that I'm most excited about. This is something that we've been semiquietly working on for the last year and a half or so. It hasn't yet been launched. It's called Duolingo. Since it hasn't been launched, shhh!
Yeah, I can trust you'll do that. So this is the project. Here's how it started. It started with me posing a question to my graduate student, Severin Hacker. OK, that's Severin Hacker. So I posed the question to my graduate student. By the way, you did hear me correctly; his last name is Hacker.
So I posed this question to him: How can we get 100 million people translating the web into every major language for free? There's a lot of things to say about this question. First of all, translating the web. Right now, the web is partitioned into multiple languages. A large fraction of it is in English. If you don't know English, you can't access it. But there's large fractions in other different languages, and if you don't know them, you can't access it. So I would like to translate all of the web, or at least most of it, into every major language. That's what I would like to do.
Now, some of you may say, why can't we use computers to translate? Machine translation is starting to translate some sentences here and there. Why can't we use it to translate the web? The problem with that is it's not yet good enough and it probably won't be for the next 15 to 20 years. It makes a lot of mistakes. Even when it doesn't, since it makes so many mistakes, you don't know whether to trust it or not.
[At often, the goat-time install a error is vomit.]
Then comes the first part of the question.
[How many times like the wind, a pole, and the dragon?]
Then comes my favorite part of the question.
[This insult to father's stones?]
And then comes the ending, which is my favorite part of the whole thing.
[Please apologize for your stupidity. There are a many thank you.]
OK, so computer translation, not yet good enough. So back to the question. So we need people to translate the whole web. So now the next question you may have is, well, why can't we just pay people to do this? We could pay professional translators to translate the whole web. We could do that. Unfortunately, it would be extremely expensive. For example, translating a tiny fraction of the whole web, Wikipedia, into one other language, Spanish. OK? Wikipedia exists in Spanish, but it's very small compared to the size of English. It's about 20 percent of the size of English. If we wanted to translate the other 80 percent into Spanish, it would cost at least 50 million dollars — and this is even at the most exploited, outsourcing country out there. So it would be very expensive. So what we want to do is, we want to get 100 million people translating the web into every major language for free.
If this is what you want to do, you quickly realize you're going to run into two big hurdles, two big obstacles. The first one is a lack of bilinguals. So I don't even know if there exists 100 million people out there using the web who are bilingual enough to help us translate. That's a big problem. The other problem you're going to run into is a lack of motivation. How are we going to motivate people to actually translate the web for free? Normally, you have to pay people to do this. So how are we going to motivate them to do it for free? When we were starting to think about this, we were blocked by these two things. But then we realized, there's a way to solve both these problems with the same solution. To kill two birds with one stone. And that is to transform language translation into something that millions of people want to do and that also helps with the problem of lack of bilinguals, and that is language education.
So it turns out that today, there are over 1.2 billion people learning a foreign language. People really want to learn a foreign language. And it's not just because they're being forced to do so in school. In the US alone, there are over five million people who have paid over $500 for software to learn a new language. So people really want to learn a new language. So what we've been working on for the last year and a half is a new website — it's called Duolingo — where the basic idea is people learn a new language for free while simultaneously translating the web. And so basically, they're learning by doing.
So the way this works is whenever you're a just a beginner, we give you very simple sentences. There's a lot of very simple sentences on the web. We give you very simple sentences along with what each word means. And as you translate them and as you see how other people translate them, you start learning the language. And as you get more advanced, we give you more complex sentences to translate. But at all times, you're learning by doing.
Now, the crazy thing about this method is that it actually really works. People are really learning a language. We're mostly done building it and now we're testing it. People really can learn a language with it. And they learn it about as well as the leading language learning software. So people really do learn a language. And not only do they learn it as well, but actually it's more interesting. Because with Duolingo, people are learning with real content. As opposed to learning with made-up sentences, people are learning with real content, which is inherently interesting. So people really do learn a language.
But perhaps more surprisingly, the translations that we get from people using the site, even though they're just beginners, the translations that we get are as accurate as those of professional language translators, which is very surprising. So let me show you one example. This is a sentence that was translated from German into English. The top is the German. The middle is an English translation that was done by a professional translator who we paid 20 cents a word for this translation. And the bottom is a translation by users of Duolingo, none of whom knew any German before they started using the site. If you can see, it's pretty much perfect. Of course, we play a trick here to make the translations as good as professional language translators. We combine the translations of multiple beginners to get the quality of a single professional translator.
Now, even though we're combining the translations, the site actually can translate pretty fast. So let me show you, this is our estimates of how fast we could translate Wikipedia from English into Spanish. Remember, this is 50 million dollars' worth of value. So if we wanted to translate Wikipedia into Spanish, we could do it in five weeks with 100,000 active users. And we could do it in about 80 hours with a million active users. Since all the projects my group has worked on so far have gotten millions of users, we're hopeful that we'll be able to translate extremely fast.
Now, the thing that I'm most excited about with Duolingo is I think this provides a fair business model for language education. So here's the thing: The current business model for language education is the student pays, and in particular, the student pays Rosetta Stone 500 dollars.
That's the current business model. The problem with this business model is that 95 percent of the world's population doesn't have 500 dollars. So it's extremely unfair towards the poor. This is totally biased towards the rich. Now, see, in Duolingo, because while you learn, you're actually creating value, you're translating stuff — which, for example, we could charge somebody for translations, so this is how we could monetize this. Since people are creating value while they're learning, they don't have to pay with their money, they pay with their time. But the magical thing here is that is time that would have had to have been spent anyways learning the language. So the nice thing about Duolingo is, I think, it provides a fair business model — one that doesn't discriminate against poor people.
So here's the site. Thank you.
We haven't yet launched, but if you go there, you can sign up to be part of our private beta, which is probably going to start in three or four weeks. We haven't yet launched it.
By the way, I'm the one talking here, but Duolingo is the work of a really awesome team,
some of whom are here. So thank you.
After re-purposing CAPTCHA so each human-typed response helps digitize books, Luis von Ahn wondered how else to use small contributions by many on the Internet for greater good. In this talk, he shares how his ambitious new project, Duolingo, will help millions learn a new language while translating the web quickly and accurately — all for free.
Luis von Ahn builds systems that combine humans and computers to solve large-scale problems that neither can solve alone.
Luis von Ahn builds systems that combine humans and computers to solve large-scale problems that neither can solve alone.