Over the past three years, the adoption of artificial intelligence by global governments has increased by over 50%, according to the study «National Large Language Models» conducted by the Ukrainian AI community.
The primary driver behind this surge is the development of local large language models that support the creation of high-quality AI solutions for sectors such as public administration, healthcare, defense, education, and agriculture.
«Advancing artificial intelligence is one of our current priorities. We are integrating AI into various projects, from education to defense tech. This enables us to become faster and more efficient as a government. Hence, our goal this year is to accelerate and fully integrate AI into our initiatives. Recently, we established the WINWIN AI Center of Excellence. One of its tasks is to launch the national LLM, which will serve as a foundation for the quality development of state products and expedite AI integration into Ukrainian businesses,» stated Mykhailo Fedorov, Vice Prime Minister for Innovation, Education, Science, and Technology — Minister of Digital Transformation of Ukraine.
A large language model (LLM) is a type of AI that functions similarly to the human brain, capable of not only analyzing texts but also generating new content. Language models form the basis for AI chatbots like ChatGPT, Gemini, and DeepSeek.
To produce text, these models are trained on vast amounts of data. National models are distinct from English-language chatbots like ChatGPT because they are based on data in the local languages of their respective countries, such as French, Bulgarian, etc.
The researchers analyzed 18 macro-regions, representing 85% of the world population, and identified key advantages of creating models in local languages.
Experiences from various countries show that large language models:
- simplify the management of regulatory documents (Albania, Serbia)
- enhance communication between authorities and citizens (Netherlands)
- optimize public procurement processes and translation of legal materials (Albania)
- improve agricultural efficiency through personalized farmer consultations (Nigeria)
- enhance medical processes by structuring medical records and predicting disease risks (Sweden, Kenya)
- assist in defense by sourcing information from open sources (OSINT) and tailoring responses to military inquiries (USA, China)
- can effectively inform citizens about government assistance (North Macedonia)
By developing national models, countries are creating AI assistants and specialized chatbots for citizens, including doctors, students, farmers, civil servants, and other professionals.
This research was conducted by the Ukrainian AI community led by the technology company OpenBabylon, lang-uk community, and the Ukrainian Catholic University.