In Ukraine, development is underway of a national large language model (LLM) called “Syaivo.” The main goal of the project is to create artificial intelligence that accurately understands the Ukrainian cultural context, history, and unique dialects. Details of the development were shared by Acting Minister of Digital Transformation Oleksandr Bornyakov.
A key stage of training the model was the collection of unique data that had not previously been used to train global chatbots. More than 50 organizations have already joined the initiative.
The largest provider of knowledge was the Ukrainian State Archives, which contributed 10 terabytes of information for the project. For comparison, this amount of data is equivalent to a library of 70,000 printed books. This is the first time in Ukraine’s history that archival funds have been directly involved in creating advanced digital services.
According to Bornyakov, Ukraine already holds leading global positions in the speed of archive digitization. However, the government plans a significant expansion of this scale. Around 150 million digital copies of documents are currently available.
It is forecast that by the end of 2026, the number of digital copies will exceed 200 million. All this data will form the training base for Syaivo.
