Kyivstar is actively developing an artificial intelligence language model, which it is working on together with the Ministry of Digital Transformation. The company is currently testing a small model, which should become the basis for creating a larger one.
Kyivstar is actively developing an artificial intelligence language model, which it is working on together with the Ministry of Digital Transformation. The company is currently testing a small model, which should become the basis for creating a larger one.
Kyivstar CEO Oleksandr Komarov told LIGA.net about this.
The model is not yet public. Komarov hopes that the project will be completed by the end of 2026.
The main reason Kyivstar considers its own model necessary is not an attempt to compete with ChatGPT or other global AI systems in scale. It is primarily about Ukrainian data and digital sovereignty.
According to Komarov, the Ministry of Digital Affairs, together with relevant state institutions and departments, has prepared a special Ukrainian dataset for training the model. «From the perspective of Ukrainianness, this will be the best AI model,» the CEO said.
Another argument is security. Komarov believes that Ukraine may need its own AI infrastructure for technologies, particularly defense ones, that use Ukrainian data.
There is also a purely economic reason. According to Komarov, a proprietary model could make token processing cheaper. This is especially important, he says, for state digital services like Diia, where using Western models on a large scale can become expensive.
