AsianFin -- Duolingo may be headquartered in Pittsburgh, but its green owl mascot is making waves far beyond American borders—particularly on Chinese social media.

On platforms like Douyin and Xiaohongshu, phrases such as "Duolingo Victims," "Loyal Green Bird Troops," and "Duolingo Aftereffects" regularly trend as users flaunt their shaky, Duolingo-learned Cantonese in videos that mix hilarity with cultural curiosity.

These clips—often limited to ordering dim sum staples like har gow and siu mai—rack up hundreds of thousands of views, underscoring how Duolingo's bite-sized, gamified lessons have turned language learning into a viral social experiment.

But behind the jokes lies a global edtech company with serious traction. According to its Q1 2025 earnings report, Duolingo boasts over 130 million monthly active users and 46.6 million daily active users worldwide. While the company hasn't disclosed exact user figures for China, the market is reportedly one of its fastest-growing.

Duolingo's popularity is largely driven by its transformation of language learning into a mobile gaming experience. The platform's game-like interface—complete with treasure chests, XP points, and league rankings—keeps users engaged through competition and rewards. "It doesn't feel like studying—it feels more like playing," said one user who's logged over 400 consecutive check-ins.

That shift has paid off. Since the introduction of its leaderboard system in late 2022, average time spent on the app has surged by 17%, with the number of highly active users—those studying at least five hours a week—doubling.

Yet, Duolingo's stickiness isn't just about streaks and coins. Behind the scenes, the company uses advanced machine learning to push personalized "nudges" and content reminders, making it harder for users to quit. One lapsed user recalled feeling "guilty" every time she saw the app's persistent owl icon on her home screen: "Eventually, I came back."

A product lead at a Chinese edtech firm likened the platform to Candy Crush, tapping into users' reward-seeking psychology. "The gameification model is addictive—it creates a low-friction learning habit," he said.

But the real engine driving Duolingo's global expansion is artificial intelligence. What began with adaptive testing and spaced repetition has evolved into a full-blown AI-first strategy.

Since 2021, Duolingo has partnered with OpenAI to integrate generative models into course development, lesson feedback, and even new features like video-based roleplay conversations. In April, CEO Luis von Ahn told employees that the company was "going all-in on AI." Shortly afterward, Duolingo launched 148 new language courses—developed entirely by AI—nearly doubling its total course count.

The cost-saving impact has been dramatic. Von Ahn noted it took 12 years to build the platform's first 100 courses; with AI, Duolingo added 148 in just one year.

However, Duolingo's reliance on generative AI has sparked criticism. On Chinese platforms and Reddit forums alike, users complain that AI-generated content often includes awkward phrasing, poor pronunciation, and illogical jumps in difficulty.

"AI can generate grammatically correct sentences," said one online English instructor, "but it struggles with nuance, context, and emotional resonance—things human teachers excel at."

Still, the business case is clear. Manually curating content is expensive. AI, by contrast, can churn out lessons at minimal marginal cost—a key advantage as Duolingo moves into new verticals like music and math.

Critics say Duolingo's fragmented, repetition-heavy method falls short for serious learners. "It's good for beginners," said a long-time user of the Japanese course. "But when you want to prepare for an exam, you still need real textbooks."

That sentiment may not concern von Ahn. When asked how the company balances educational value with engagement, he responded: "It's simple—always choose engagement."