“Duolingo may generate, record, and store audio recordings or transcripts of the text and audio you submit, and use these recordings or transcripts for product improvement and personalization purposes, including training and running Duolingo's own artificial intelligence models.”
This analysis describes what Duolingo's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
The clause permits Duolingo to use user-submitted content to train proprietary AI models, meaning user interactions directly feed Duolingo's AI development.
The updated policy now discloses that audio collection for product improvement applies to all users across all platforms (iOS, Android, and web), whereas previously the policy stated audio was collected only from iOS users. Android and web users are now subject to audio collection for product improvement purposes. All users can opt out of audio sharing within app Settings.
View change record →The updated policy introduces a new Session Replay service that logs user activity alongside FullStory, capturing clicks, mouse movements, scrolling, typing, device information, navigation, and learning activity. Both services can be disabled using the Tracking toggle in app Settings, and both are automatically disabled for child users. The new Math Tutor feature shares audio with Apple for speech recognition and immediate deletion, but retains and shares text transcripts with AI vendors. For payment processing and fraud prevention, Duolingo may retain IP addresses beyond the standard 30-day period. You can disable both FullStory and Session Replay activity recording by toggling the Tracking setting within the app.
View change record →Text and audio you submit to Duolingo may be recorded, stored, and used to train Duolingo's AI systems and personalize your experience.
How other platforms handle this
The personal information we collect may be stored and processed in your country or region, or in any other country where we or our affiliates, subsidiaries, or vendors maintain facilities.
ALPR Information is stored for approximately 30 days before it is permanently deleted unless the information is retained for investigative or legal purposes.
if you choose to voluntarily submit a video to help improve our Services, we store this person detection information for as long as needed...but in no event longer than 3 years after the date you submitted the video.
"Duolingo may generate, record, and store audio recordings or transcripts of the text and audio you submit, and use these recordings or transcripts for product improvement and personalization purposes, including training and running Duolingo's own artificial intelligence models.Excerpt from Duolingo's Privacy Policy
Get the weekly coverage letter
Companies change their terms without announcing it. We capture governance documents, compare each new version with the last, and publish what changed, with the source text behind it. One email a week on what we found and what we could not reach.
The clause permits Duolingo to use user-submitted content to train proprietary AI models, meaning user interactions directly feed Duolingo's AI development.
Text and audio you submit to Duolingo may be recorded, stored, and used to train Duolingo's AI systems and personalize your experience.
ConductAtlas has identified this type of provision across 267 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Duolingo.