The Cabinet of Ministers of Ukraine has adopted a resolution on the creation of a national data corpus designed for training and testing large language models (LLMs). The document, reported by RBC-Ukraine citing the text of the resolution, establishes the state's role as the curator of a unified database of text resources in the Ukrainian language and sets clear access rules for artificial intelligence developers. The deployment of the platform will be handled by the Ministry of Digital Transformation jointly with the state-owned enterprise "DIYA", and developers will be able to use the collected dataset for free based on a license agreement.
Role Distribution and System Development Timelines
The resolution defines a clear division of functions between project participants. The Ministry of Digital Transformation acts as the organizer and curator of the national database, while the SOE "DIYA" is responsible for technical filling, data structuring, and cybersecurity. After the platform's launch, the ministry, jointly with the Ministry of Education and Science and with the participation of the National Academy of Sciences, is given six months to develop proposals for the further development of the national language resource system. Thus, the launch of the data corpus is viewed not as a final stage, but as a starting point for a multi-year process of forming the country's digital linguistic infrastructure.
What Will Be Included in the National Corpus
The national corpus will combine anonymized structured and unstructured texts in the Ukrainian language that reflect the cultural and historical context of the state. Its composition will include state databases, literary works, scientific papers, and archival materials. The resolution provides two replenishment mechanisms: the mandatory transfer of open datasets by executive authorities within a month after receiving a request from the Ministry of Digital Transformation or SOE "DIYA", and the voluntary participation of private companies, libraries, educational and scientific institutions, individual entrepreneurs, and citizens based on license agreements in compliance with intellectual property legislation.
Who Will Get Access and Under What Conditions
Developers will be able to use the collected dataset for free, but the circle of applicants is strictly limited. Access will be provided only to Ukrainian legal entities and individual entrepreneurs who have confirmed experience in creating or testing LLMs and commit to using the materials exclusively for these tasks. SOE "DIYA" will review applications from developers within ten working days. If the system records a violation of security rules or misuse of information, access is cancelled instantly.
Strict Security Safeguards and Protection of Rights Holders
The resolution establishes a number of prohibitions aimed at data protection. Access to the database is closed to financial institutions, investment funds, sanctioned companies, as well as businesses operating in temporarily occupied territories, Russia, or Belarus, or having ultimate beneficiaries or shareholders with a stake of 10 percent or more. A separate mechanism for protecting rights holders is prescribed: upon receipt of a complaint about the unlawful use of text, access to the disputed dataset is temporarily blocked within one working day for the duration of the verification. These measures are designed to balance the interests of developing the domestic AI sector and protecting copyright, as well as to prevent the use of the national database in the interests of third countries.
Significance of the Project for the Ukrainian AI Sector
The creation of the national data corpus is a systematic step in forming a sovereign artificial intelligence ecosystem in the Ukrainian language. Free access to structured and unstructured texts combining literary, scientific, and archival heritage lowers the entry barrier for local LLM developers and creates the basis for generating specialized language models that adequately reflect Ukrainian linguistics and cultural context. The joint work of the Ministry of Digital Transformation, SOE "DIYA", MES, and the NAS of Ukraine over the next six months should determine a roadmap for the further expansion and development of the national language resource system.
