ASPECTS OF IMPLEMENTING ARTIFICIAL INTELLIGENCE IN THE PUBLIC ADMINISTRATION SYSTEM OF UKRAINE

Abstract

The article presents a comprehensive analysis of the theoretical and
practical aspects of integrating artificial intelligence into the public administration
system of Ukraine. The study is structured around examining transformational processes
through the lens of three interconnected dimensions: the current digital readiness of the
state, international experience in utilizing large language models (LLMs) in the public
sector, and the processes of shaping an adaptive regulatory environment. It is
demonstrated that such a comprehensive approach allows for the formation of a clear
understanding of the architecture of intelligent governance.
It is shown that Ukraine is actively developing its own digital governance
ecosystem based on the core platforms Diia, Prozorro, and Trembita. Within this system,
a systematic integration of intelligent AI assistants for public service delivery and citizen
communication is taking place, alongside automated tools for the analysis and legal
review of regulatory acts. The study substantiates that implementing the bottom-up
approach to artificial intelligence regulation, as outlined in the respective national
roadmap, enables a gradual preparation of private businesses, state institutions, and
society for alignment with the EU AI Act requirements, thereby significantly minimizing
potential operational risks.
Based on a comparative analysis of national LLM initiatives in Bulgaria, Greece,
the Netherlands, Sweden, Singapore, and Albania, specific strategic recommendations
are formulated to strengthen Ukraine's institutional capacity in the technological domain
of artificial intelligence. The findings of the study identify and classify key challenges that
hinder the scaling of innovations within the government sector. These include an acute
shortage of qualified personnel, persistent cyber threats to critical infrastructure under
martial law, the presence of legacy IT systems, and the risks of algorithmic bias in
decision-making processes.

References
1. Organisation for Economic Co-operation and Development. (2019). OECD
principles on artificial intelligence. Retrieved from https://oecd.ai/en/aiprinciples [In English].
2. Hoekstra, M., Dom, L., & van Veenstra, A. F. (2025). Quickscan AI in de publieke
dienstverlening III. TNO. Retrieved from
https://vector.tno.nl/publish/pages/5406/samenvatting_quickscan_ai.pdf
[In Dutch].
3. Natsionalni velyki movni modeli: Tsinnist, stratehii, rehuliuvannia [National large
language models: Value, strategies, regulation]. (2025). Retrieved from
https://surl.li/qegcsy [In Ukrainian].
4. Cabinet of Ministers of Ukraine. (2021). National strategy for the development of
artificial intelligence in Ukraine for 2021–2030 (Order No. 1556-r). Retrieved
from https://zakon.rada.gov.ua/laws/show/1556-2021-%D1%80 [In Ukrainian].
5. V Ukraini rozrobleno dorozhniu kartu z rehuliuvannia shtuchnoho intelektu
[A roadmap for artificial intelligence regulation has been developed in Ukraine].
(2025). Retrieved from https://news.online.ua/ukrayina-maje-ocoliti-trend-si-usviti-fedorov-865399/ [In Ukrainian].
6. Ministry of Digital Transformation of Ukraine. (2025). Diia AI-asystent
“Natalka”. Retrieved from https://diia.gov.ua [In Ukrainian].
7. Dorozhnia karta z rehuliuvannia shtuchnoho intelektu v Ukraini [Roadmap for
artificial intelligence regulation in Ukraine]. (2025). Retrieved from
https://surl.li/akdexf [In Ukrainian].
8. Fedorov, M. (2025). Quote of Minister Mykhailo Fedorov. Retrieved from
https://surl.li/udeklf [In Ukrainian].
9. Ministry of Digital Transformation of Ukraine. (2025). Bila knyha z pytan
rehuliuvannia shtuchnoho intelektu [White paper on artificial intelligence
regulation]. Retrieved from https://surl.li/akdexf [In Ukrainian].
10. Dorozhnia karta z rehuliuvannia shtuchnoho intelektu v Ukraini: Rozrobka ta
steikkholdery [Roadmap for artificial intelligence regulation in Ukraine:
Development and stakeholders]. (2025). Retrieved from https://surl.li/hchfmv
[In Ukrainian].
11. Bulgaria leads Europe with first state-supported AI language model. (2025).
Retrieved from https://surl.lu/gyxatl [In English].
12. Meltemi: The first Greek large language model. (2024). Retrieved from
https://surl.li/bjrhnw [In English].
13. Taylor, A. (2024). Albania: Pioneering AI, lagging in legal protections.
Respublica. Retrieved from https://respublica.edu.mk/blog-en/politics/albaniapioneering-ai-lagging-in-legal-protection [In English].
14. COMtext.SR – ICEF. (2018). Retrieved from
https://www.ic.etf.bg.ac.rs/projects/comtext-sr/ [In English].
15. North Atlantic Coalition. (2024). Codi, the virtual policy assistant. Retrieved from
https://surl.li/gtbiyr [In English].
16. AI Sweden. (2023). Open release of the first large Nordic language model GPTSW3. Retrieved from https://surl.li/nzjnyd [In English].
17. Introducing Artificial Intelligence AI@NIE (AI@NIE). (2022). Retrieved from
https://surl.lu/gnjtxc [In English].
18. BharatGen. (2024). About. Retrieved from https://bharatgen.tech/about.html
[In English].