RESCUING THE UZBEK LANGUAGE FROM DIGITAL EXTINCTION: A COMPUTATIONAL FRAMEWORK BASED ON TRANSFORMER ARCHITECTURES

Abstract

The digital divide between globally dominant languages and low-resource languages threatens the plurality of language cultures and languages. As AI and NLP are becoming the go-to interfaces between information exchanges, languages with inadequate computational architectures are likely to experience “digital extinction”. One prominent example of this kind of situation is the Uzbek language, whose millions of speakers face such a crisis today because its complex agglutinative morphology and a lack of high-quality annotated datasets are largely missing in such systems. This article outlines a strategic framework that combines philological expertise with contemporary Transformer-based architectures. In this way, this research seeks to reduce semantic errors in AI processing by at least 30% by establishing a dedicated morphological analyzer and a robust digital corpus, to make sure that the Uzbek language can remain a critical tool for science, innovation, and global discourse in the digital era.