A national database for artificial intelligence training will be created in Ukraine, - the decision of the Cabinet of Ministers.
The government mandated the Ministry of Digital Transformation to create a Ukrainian data corpus within three months for the training and testing of large language models (LLM).
According to Resolution No. 1,416 of November 11, executive authorities must, at the request of the Ministry for Digital Development or the State Dia Enterprise, transfer structured and unstructured anonymous data within a month.
The transfer of information must be carried out in accordance with intellectual property laws.
It is also recommended to involve scientific institutions, archives, libraries, educational institutions, enterprises and private individuals in the creation of the database.
Companies and entrepreneurs who develop, train or test language models will be able to access the data. The information obtained may be used exclusively for these purposes.
At the same time, access to companies and entrepreneurs operating in Russia, Belarus and countries that are part of customs or military unions with them, as well as to people under sanctions, will be prohibited. Restrictions also apply to certain categories of citizens and legal entities of aggressive states.
Requests for access will be reviewed by the government enterprise Dea within 10 working days.
Also, the Ministry for Digital Development, together with the Ministry of Education and Science and in cooperation with the National Academy of Sciences, must prepare proposals for the further development of the national system of language resources.