The Business and Technology University has launched a Georgian language digital sovereignty project. The project aims to ensure that the Georgian language does not lose its semantic accuracy, grammatical logic and cultural depth in the era of artificial intelligence. BTU’s initiative is aimed at ensuring that AI systems not only read or translate text into Georgian, but also correctly perceive the content and context, grammatical norms and rules, idioms and expressions, and are able to analyze the text in terms of value.

“Language is not just a sequence of words. The meaning is read in verbs, cases, word order, context and phrases that cannot be translated literally. Therefore, teaching Georgian correctly to AI does not just imply accumulating texts, it also requires an understanding of the internal structure of the language,” BTU researchers explain.

Within the framework of the project, BTU has already created an initial architecture that can become the basis of a full-scale digital infrastructure of the Georgian language. The code for Georgian verbs and complex verb forms, a database of natural Georgian sentences, AI error types, and an initial model for special AI testing for the Georgian language have been developed.

As stated in BTU, the digital sovereignty of the Georgian language is a task of national importance in the 21st century. If AI understands Georgian superficially, errors may appear in education, media, business, legal texts, public services and everyday communication. If the Georgian language is properly modeled, the country will have better Georgian-speaking AI, more accurate translation, better digital services, powerful educational resources and a reliable digital perception of Georgian culture.

The project is important not only for the preservation of the language, but also for Georgia's technological development, education, business, media and digital independence. In BTU's view, the Georgian language in the AI ​​era should be presented as a system with its own grammatical, semantic, cultural and informational logic.

(AD)