Google's AI photo app uses crowdsourcing to preserve endangered languages

Woolaroo lets cultures protect their vocabularies on their own terms.


Google has a new way to preserve endangered languages: give cultures the AI tools they need to protect the languages themselves. The company has launched Woolaroo, an open source photo translation web app (also available through Google Arts & Culture for Android and iOS) that uses machine learning and image recognition to help preserve languages on the brink. As a user, you just have to point your phone's camera at an object to have the AI recognize and describe it in a given language, complete with pronunciation.

Woolaroo's real power comes from its open nature, however. Communities can use the platform to expand vocabularies on their own terms. If you remember a word that hasn't been covered yet, you can add it (and its pronunciation) with relatively little trouble. That could be particularly important for languages that don't have unique words to describe modern concepts like phones or computers. You can modify or delete entries if they're inaccurate, too.

The app initially offers exploration of 10 languages from around the world, including Maori, Yiddish and Australia's indigenous Yugambeh. We won't be surprised if that expands quickly, though. UNESCO has determined that "at least" 2,572 of the world's estimated 6,000 languages face at least some danger. Although Woolaroo won't guarantee that languages will stay in active use, it could keep them and their associated histories from fading into obscurity.

This article contains affiliate links; if you click such a link and make a purchase, we may earn a commission.