Transforming Learning into an Irresistible Experience

Hi there to all of you! Even though not everyone is immediately familiar with Guatemala, that is where I am from. Just to clarify, Guatemala is not where prisoners are kept—that's Guantanamo. For those in the US, Guatemala may be thought of as "Mexico's Mexico," as it is situated directly beneath Mexico. Similar to the US's concerns regarding illegal immigration from Mexico, Mexico also harbors similar worries regarding illegal immigration from Guatemala. Although the nation is smaller and less developed, the Mexican cuisine is surprisingly good.

Guatemala, as a nation, faces significant challenges with poverty. Education is often touted as a means to bridge social and economic gaps. However, in practice, it can sometimes widen the divide. Wealthier individuals can afford top-notch education and thus maintain their financial advantage, while those less fortunate may struggle with basic literacy, perpetuating the cycle of poverty. This disparity is especially pronounced in poorer countries.

Even though I did not have a wealthy upbringing, I was fortunate enough to get an education that is normally only given to the wealthy. Raised as an only child by a single mother, she devoted her life to my education, setting the path for my studies in the United States and ultimately earning me a doctorate in computer science.

About ten years ago, I made the decision to address educational inequality. Together, Severin, my PhD candidate, and I at Carnegie Mellon University's Department of Computer Science tackled this issue. We decided to start with a specific topic instead of attempting to cover the vast array of topics covered in education as a whole because we had to concentrate more tightly.

Initially, we considered teaching math or computer science—both areas we are passionate about. However, after careful consideration, we decided to focus on teaching foreign languages. This might seem surprising, but there are compelling reasons behind this choice.

Firstly, there is a massive global demand for language learning. Approximately two billion people are currently learning a foreign language, with a significant majority focusing on English. Knowledge of English, in particular, can profoundly enhance one’s earning potential across many countries. This contrasts with subjects like math, where knowledge alone doesn’t directly increase income potential. For instance, learning math is a step towards learning physics, which then leads to becoming an engineer—a more indirect route to higher earnings.

Given this, we concluded that teaching languages was the most impactful starting point. Moreover, we realized that reaching a broad audience effectively required leveraging mobile technology. Building schools worldwide would be prohibitively expensive, but smartphones are becoming increasingly ubiquitous. Thus, we decided to create a mobile app that makes learning foreign languages accessible to everyone. That app is Duolingo.

Duolingo operates on a freemium model. This means users can access the app and learn as much as they want without any cost, though they might encounter ads. Users can also opt to pay a subscription fee to remove ads. Interestingly, the majority of Duolingo’s revenue comes from these subscriptions, typically from users in wealthier countries. This setup results in a form of wealth redistribution, where users in affluent countries essentially subsidize the education of those in less wealthy nations.

However, delivering education through smartphones presents its own challenges. Smartphones are packed with incredibly engaging and addictive apps—TikTok, Instagram, and mobile games—which can make educational content seem less appealing by comparison. To overcome this, we had to make learning as engaging as possible.

Duolingo employs various techniques to enhance user engagement, borrowing strategies from social media and gaming. One effective method is the “streak” feature, which tracks and displays the number of consecutive days a user has engaged with the app. This creates a psychological commitment to maintain the streak, encouraging daily practice. We have over three million daily users, with streaks lasting more than a year.

Another engagement tool is notifications. Our AI system sends personalized reminders to encourage users to return to their lessons, with the most effective time being 24 hours after their last session. Interestingly, if users stop using the app for a week, notifications cease, but a special message is sent to encourage their return, leveraging a bit of passive aggression to nudge them back.

Our green owl mascot, who sends these notifications, has become a cultural icon, inspiring countless memes and humor. The playful approach has helped Duolingo integrate into popular culture, showing that educational tools can be both effective and entertaining.

While I believe it’s challenging to make educational apps as engaging as entertainment platforms, the meaningfulness of learning can provide intrinsic motivation. If we can make educational content engaging, even if it’s not as captivating as TikTok, we can still reach millions. For instance, Duolingo has more language learners in the US than all US high schools combined.

My hope is that we can apply the same principles used in Duolingo to other subjects like math and science, leveraging mobile technology to provide high-quality education to everyone, regardless of their socioeconomic status. Imagine a future where screen time is used for enriching learning experiences rather than just entertainment.

Thank you very much.

Helen Walters: Luis, thank you for that insightful talk. Could you elaborate on how you envision applying these ideas to other subjects, such as math or science?

Luis von Ahn: Certainly. For subjects that involve repetitive practice, such as basic math or reading, gamification can be particularly effective. These subjects lend themselves well to repetitive exercises, making them ideal candidates for a model similar to Duolingo. However, for subjects requiring more explanation, such as complex theories in physics, high-quality video content might be more appropriate. But for anything that can be learned through repetition, applying these techniques can make learning both engaging and effective.