Duolingo collects extensive data to power its gamification and AI systems. Voice recordings, detailed learning patterns, and behavioral analytics create comprehensive profiles of how you learn. This data trains Duolingo's AI, informs product development, and supports the advertising-based free tier. Many users are children, raising additional concerns about data collection.
| Data Type | Collected | Shared | Sold |
|---|---|---|---|
| Voice Recordings | Yes | AI Development | No |
| Learning Patterns | Yes | Research Partners | No |
| Behavioral Data | Yes | Analytics Platforms | Advertising |
| Device Information | Yes | Ad Networks | Targeting |
| Error Patterns | Yes | Product Development | No |
Speaking exercises collect voice recordings that are retained and analyzed. Voice data can reveal identity, accent, age, and even health conditions. This sensitive data is used for AI training and product improvement.
Every interaction is tracked: time between answers, error patterns, session length, and engagement with gamification elements. This data powers the psychological manipulation that drives engagement.
Free tier users are the product. Behavioral data, device information, and usage patterns are shared with advertising networks to target ads both within and outside the app.
Duolingo is popular with children learning languages for school. Despite this, data collection practices are largely the same for all users, raising COPPA and child privacy concerns.