A large-scale AI model for recognizing Turkic languages has been developed in Kazakhstan.

In Kazakhstan, one of the largest artificial intelligence models in the region for automatic speech recognition in Turkic languages was presented. The development was carried out by the Kazakh startup Cybernet AI with the support of Astana Hub and the Ministry of Artificial Intelligence and Digital Development of the Republic of Kazakhstan. The project was implemented using Microsoft's GPU infrastructure.

The key feature of the solution is that the model was initially developed taking into account the linguistic features of the Turkic language group. Unlike most existing systems, which are created based on English-language data and then adapted to other languages, the new model was trained specifically for regional languages and the speech of real users.

The AI system can recognize Kazakh, Turkish, Uzbek, Kyrgyz, Azerbaijani, and Tatar languages. It also works correctly with mixed speech, which combines Turkic languages and Russian — a format widely used in everyday communication, business, and when contacting government services.

At Astana Hub, they note that launching such technology creates conditions for scaling digital services in the native languages of the countries of the Turkic world. According to representatives of the ecosystem, this makes it possible to increase the accessibility of AI solutions for business and government, as well as reduce dependence on foreign technologies.

Developers emphasize that the model is resistant to background noise, various accents, and language switching within a single dialogue. This makes it applicable for contact centers, banks, telecom companies, service organizations, and government structures, where large volumes of voice calls are processed daily.

At Cybernet AI, they state that the project is a contribution to the development of the local AI industry and the technological sovereignty of the region. Creating own language models allows taking into account the cultural and speech features of users and improving the quality of digital services.

Source: Gov.kz