Yugra Researchers Train AI to Support Indigenous Languages of the North
Experts at the 17th International IT Forum in Khanty-Mansiysk explored how digital technologies can help preserve the endangered languages of Indigenous peoples, highlighting AI-powered language preservation initiatives in Russia's Yugra region.
A key focus of the discussion was the work of the Yugra Research Institute of Information Technologies. Over the past three years, researchers have digitized 400 books – nearly every written source held in the regional library – and built a language corpus containing 310,000 translation pairs. As a result, one of the Indigenous languages of the North has been added to Yandex's online translator.
Experts emphasized that a language rapidly loses its chances of survival if it has no presence in the digital world. Training AI models requires digitized texts, recordings of spoken language, and images. Forum participants agreed that "artificial intelligence technologies should not threaten cultural diversity." Instead, they said, AI should help preserve humanity's unique cultural heritage.
The Khanty-Mansi Autonomous Okrug is also developing a language corpus for Khanty that includes four dialects. The project is scheduled for completion by Nov. 30.
