This chapter examines Ukraine’s efforts to adopt artificial intelligence (AI) in government and provides targeted policy recommendations to support its responsible and effective use for policy design, service delivery and the public sector. Applying the OECD Framework for Trustworthy AI in Government, the chapter assesses Ukraine’s progress across four pillars: enablers, guardrails, engagement, and impact monitoring. It builds on Ukraine’s strong digital transformation and identifies opportunities to further strengthen its AI governance. Drawing from international practices, the chapter focuses on guiding Ukraine in making effective use of AI to enhance public service design and delivery with sequenced actions for the immediate term, recovery and reconstruction, and long-term development in line with global standards.
Public Governance Review of Ukraine
Abstract
Artificial intelligence (AI) has the potential to improve public sector productivity, responsiveness, and accountability. When used in a trustworthy manner, AI can support the government in automating routine processes, tailoring services to individual needs, supporting better decision-making, detecting fraud or anomalies, and enhancing the work of public servants. As outlined in the OECD report Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Function (OECD, 2025[1]), AI capabilities can translate into tangible benefits for governments, including faster service delivery, cost savings, and enhanced public satisfaction. Across OECD countries and beyond, AI is increasingly being used in public services, civic participation, justice administration and other functions, particularly to improve end-to-end service delivery.
However, realising the benefits of AI in government requires addressing implementation challenges and managing potential risks. AI systems can introduce new challenges and reinforce existing ones, including issues arising from skewed data, limited transparency and explainability, and reduced human oversight due to over-reliance on automated systems. If not adequately addressed, these challenges may lead to unintended outcomes for individuals and groups, potentially weakening public trust. In addition, many AI initiatives in government remain at the pilot stage, reflecting difficulties in scaling solutions, challenges in measuring and assessing impact, gaps in data and data governance, limited investment in skills, and the absence of clear governance arrangements, including guidelines and monitoring mechanisms.
At the same time, the risks of inaction should not be overlooked. Delays in adopting AI and translating available capabilities into practice can result in avoidable financial and non-financial costs, while limiting improvements in service quality, efficiency, and productivity. They may also contribute to a widening gap between public and private sector capabilities and constrain governments’ ability to effectively oversee evolving technologies. In some cases, uncertainty and risk-averse approaches may further slow adoption, even for lower-risk, high-benefit uses, limiting the extent to which governments can realise AI’s potential.
To address these issues, governments must build an environment where they can use AI responsibly and ethically. Based on the OECD Framework for Trustworthy AI in Government, this includes developing enablers such as governance, digital infrastructure, quality data, and skilled talent; establishing guardrails through policies, oversight bodies, and transparency measures; ensuring engagement to bring a range of stakeholders into shaping the AI agenda and developing AI solutions; and anticipating and monitoring the impact of trustworthy use of AI systems in government (OECD, 2024[2]). While many efforts are still emerging, early lessons can guide future government actions in using AI to create more responsive and effective public services that are available to all groups in society.
Ukraine has made substantial advancements in the digital transformation of the public sector. This has demonstrated resilience and innovation in challenging circumstances due to Russia’s war of aggression against Ukraine. Driven by the strategic leadership of the Ministry of Digital Transformation (MDT) and its network of Chief Digital Transformation Officers (CDTOs), the country has rapidly advanced its digital ecosystem, demonstrating that government transformation can reinforce both service continuity and national resilience. Initiatives such as the State in a Smartphone programme, Diia, and the government interoperability platform, Trembita, have significantly enhanced the accessibility and efficiency of government services, serving over 22 million users, more than half of Ukraine's population (OECD, 2024[3]; Government of Ukraine, 2025[4]). Ukraine’s digital transformation journey has laid the groundwork for the effective use of AI in government to further enhance the design and delivery of public services.
This chapter applies the OECD Framework for Trustworthy AI in Government to Ukraine’s current AI-related policies, practices, and use cases. It assesses Ukraine’s progress and sets out policy recommendations sequenced across: (i) immediate priorities feasible under current wartime and institutional conditions; (ii) recovery and reconstruction priorities to pursue as capacity expands during rebuilding; and (iii) long-term development priorities to further align with OECD standards. Grounded in relevant international best practices, the recommendations aim to support Ukraine’s responsible and effective use of AI in government. The analysis reflects developments up to June 2025, which serves as the cut-off date.
Robust governance frameworks for AI are imperative to ensure the strategic, effective, responsible, and trustworthy use of AI in government. As governments increasingly integrate AI into their internal processes and public service design and delivery, developing clear governance structures and arrangements becomes crucial to address potential ethical, legal and operational risks. It will also support governments in ensuring alignment with relevant international regulatory standards, including, for example, the EU AI Act (Box 6.1), which lays comprehensive requirements for transparency, accountability, risk assessment and human oversight of AI systems.
The European Union (EU) AI Act is a regulation on AI that entered into force in August 2024. The regulation establishes obligations for AI based on its potential risks and level of impact. The Act identifies four different levels of risks that are relevant for governments’ use of AI.
Unacceptable risk: AI uses under this category are prohibited by the AI Act. Examples include predictive policing, ‘real-time’ remote biometric identification (including facial recognition) in publicly accessible spaces for law enforcement, social scoring, or assessing the risk of an individual committing criminal offences. Law enforcement and justice administration are among the functions of government most closely tied to this category, although some exceptions apply, such as use cases involving national security and those remaining subject to judicial oversight.
High-risk: AI uses under this category are allowed but regulated due to their significant potential harm to health, safety, fundamental rights, environment, democracy, and the rule of law. Due to its potential impact on these aspects, most government uses of AI might fall under this category. Examples include systems used to influence the outcome of elections and voter behaviour, automated processing of personal data to assess various aspects of a person’s life, assessing eligibility for benefits and services, and safety components used in the management and operation of critical infrastructure. Obligations include establishing a risk management system, conducting data governance, having in place technical documentation to demonstrate compliance, and mandatory fundamental rights impact assessments, among others.
Limited risk: These systems might include systems intended to communicate with individuals, such as chatbots, as well as systems that generate content such as text and images. These types of systems have transparency obligations where developers and deployers must ensure that end-users are aware that they are interacting with AI.
Minimal risk: These systems are unregulated, but a code of conduct is suggested. Examples include video games and spam filters.
Source: (OECD, 2025[1])
The OECD Framework for Trustworthy AI in Government
The OECD Framework for Trustworthy AI in Government has been developed to support governments in addressing key policy and governance challenges in their use of AI. The framework aims to provide a comprehensive approach to navigating the complexities of AI use in government, ensuring that AI is used in a strategic, trustworthy and efficient manner, and to support alignment with the OECD AI Principles (see more in Box 6.8) to which Ukraine has adhered. It was developed to address three significant policy questions, including:
What concrete policy actions and tools can governments develop to address existing challenges for a trustworthy use of AI in government?
Who should governments engage when developing and implementing the enablers and guardrails for the trustworthy use of AI in government?
What impact do governments strive to achieve when using trustworthy AI?
The framework calls for strengthening enablers (Pillar 1), establishing guardrails (Pillar 2), and engagement with a broad range of stakeholders during both the design and implementation stages of AI projects (Pillar 3) (Figure 6.1).
The figure below shows the pillars and dimensions for the trustworthy use of AI in the public sector.
Source: (OECD, 2025[1])
These pillars serve as the basis for analysing AI readiness in Ukraine’s public sector. The intent is to use the framework to provide Ukraine with concrete actions that can be taken to focus its efforts going forward, as well as the relevant international practices.
What enablers are in place for the use of AI in government?
The enablers are the foundational conditions and policy tools that should ideally be in place to create an enabling environment and unlock the full-scale adoption of AI in government. These include governance, capabilities (e.g. infrastructure, data, skills, and talent), collaborations and partnerships, including through public procurement. Together with the guardrails, these enablers also help mitigate risks and address key challenges to effective and trustworthy AI use in government. Nevertheless, not all enablers need to be in place from the outset; many can be developed iteratively as an approach to AI matures.
Progress to date
Ukraine has made significant progress in establishing the enablers for AI in the public sector, despite ongoing challenges. From a strategic standpoint, Ukraine has formulated strategic documents, namely the Concept of Artificial Intelligence Development in Ukraine (2020) (Government of Ukraine, 2020[5]), the Strategy for the Digital Development of Innovations in Ukraine for the period until 2030 (2024) (also known as Digital Innovation Development Strategy until 2030 – WINWIN) (Government of Ukraine, 2024[6]) and the White Paper on AI Regulation (2024) (Government of Ukraine, 2024[7]) (see more details of strategic documents in Table 6.2), which outlines both a long-term vision and short-term plans for key priority areas. Ukraine has adopted a collaborative approach in developing these documents, with contributions from various disciplines at both national and international levels. A comprehensive strategic document, the Strategy for the Digital Development of Innovations in Ukraine for the period until 2030 (2024) aims to create new opportunities for Ukrainians in the digital innovation ecosystem. The Strategy has a specific goal on AI (see more in Table 6.2), focusing on the development of internal infrastructure for research, innovation and implementation of solutions in the field of AI. In addition, the establishment of the Ministry of Digital Transformation (MDT) has been instrumental in providing the necessary leadership to drive digital transformation in Ukraine, including the adoption of AI in government.
Furthermore, the government has initiated measures to strengthen the enablers to translate its strategic approach into tangible actions. Ukraine has made significant efforts to continue building key infrastructure that enables the delivery of government services through digital channels, as well as to share and reuse data in a secure and trustworthy manner. Initiatives such as the AI/Blockchain sandbox, for both the private and public sectors, and the WINWIN AI Centre of Excellence (see Box 6.3) have helped translate concepts into use cases in line with international norms and standards. The use of common AI tools, such as government chatbots and virtual assistants, is also gaining traction, promising economies of scale and ensuring coherent AI deployment across the public sector. Table 6.1 outlines some of Ukraine’s efforts in building the enablers for the use of AI in the public sector.
The objective of Goal 16 is to consolidate Ukraine’s standing on the global stage in the area of AI, to promote collaborative efforts in technological development, and to encourage the growth and advancement of the domestic innovation ecosystem. On this note, it aims to create an enabling environment for the implementation of AI solutions, including internal infrastructure, research and innovation.
Key projects and responsible institutions
1. AI and Governance Innovation Cluster (2025-2027)
Responsible entities: MDT, Ministry of Education and Science (MES), Ministry of Economy (ME), Ministry of Energy (MEV), National Academy of Sciences (NASU), Higher Education Institutions
Objective: Establish an innovation cluster connecting research institutes, universities, startups, and private companies to jointly develop governance-focused AI solutions.
Expected outcomes: Strengthened international leadership, increased innovation capacity, and enhanced technological collaboration and exchange.
2. Ukrainian Language Resources Enrichment (2025-2027)
Responsible entities: MES, Commissioner for Protection of the State Language
Objective: Support projects enhancing the corpus of the Ukrainian language to promote linguistic research.
Expected outcomes: Improved language resources, increased accessibility, and enhanced international promotion and usage of Ukrainian.
3. White Paper on AI Regulation (2025-2027) – published
Responsible entity: Ministry of Digital Transformation (MDT)
Expected outcomes: Clear regulatory framework, enhanced predictability, and increased transparency in AI policy and implementation.
4. High-Tech AI and Blockchain (Q2, 2025)
Responsible entities: MDT, MES, Ministry of Health (MoH), Ministry of Defence (MoD), Ministry of Agrarian Policy and Food (MAPFU), Commissioner for Human Rights
Objective: Establish regulatory conditions supporting the implementation and competitiveness of innovative AI and blockchain products.
Expected outcomes: Enhanced technological innovation, increased market competitiveness, and facilitated testing and deployment of new AI-driven services.
5. AI Integration into Priority Economic Sectors (2025-2027)
Responsible entities: MDT, Ministry of Strategic Industry (MSPU), MES, ME
Objective: Approve and implement the State Targeted Scientific and Technical Program for integrating AI into priority sectors of Ukraine’s economy.
Expected outcomes: Enhanced AI-driven innovation across priority economic sectors, improved economic performance, and increased efficiency.
6. AI Competence Centre for Energy Security (Q4, 2025)
Responsible entities: MEV, MDT, National Commission for Energy and Utilities, National Power Company "Ukrenergo", Higher Education Institutions
Objective: Create a dedicated AI competence centre to enhance energy security, optimise dispatching processes, and advance "smart grid" technologies.
Expected outcomes: Integration of advanced AI in the energy sector, increased sector efficiency and reliability, reduced energy losses, improved energy supply quality, and alignment with Ukraine’s “smart grids” strategy.
Source: (Government of Ukraine, 2024[6])
In February 2025, the Ministry of Digital Transformation launched the WINWIN AI Centre of Excellence, a dedicated hub for developing and integrating AI in the public sector. The initiative is part of Ukraine’s Digital Innovation Development Strategy until 2030 (WINWIN), which aims to modernise public services and strengthen national competitiveness. The Centre is housed within the Ministry of Digital Transformation, ensuring close alignment with the national digital and AI strategies. It is led by the Chief AI Officer (CAIO) at the Deputy Minister level and the Chief Technology Officer (CTO).
Objectives and strategic role:
Advancing AI adoption in public services: The Centre is designed as a catalyst to embed AI in core public sector functions, improving citizen services and internal efficiencies. A primary objective is to integrate AI into government processes and key sectors such as defence, healthcare, education and business. Through this, the government seeks to transform service delivery, personalise services and streamline administrative tasks.
Public sector innovation and AI Governance: The Centre was developed to serve as an innovation hub for the public sector, introducing a dedicated Chief AI Officer (CAIO) role in government and a team to spearhead AI initiatives. The Centre also has a mission to unite experts, resources and innovations for AI-driven transformation of public services.
Private sector collaboration and ecosystem building: The Centre was conceived as a bridge between the public sector and the tech industry, with a key objective to foster public-private collaboration. The Centre will work with businesses, startups and academia to co-create AI systems for public needs.
The table below assesses the enablers that should be in place to strengthen the use of AI in government
Where to focus next
Despite promising progress, to leverage the full potential of AI and scale-up the ongoing efforts, Ukraine can do more to institutionalise AI governance arrangements, enhance its data and digital infrastructure, and strengthen monitoring of trustworthy AI use in the public sector.
A critical challenge for the future is to formalise Ukraine’s governance arrangements around AI across government. Building on the strong leadership of the MDT and the recently created role of the Chief AI Officer (CAIO), there is a growing need from the public sector institutions and partners for a permanent coordination mechanism at the centre dedicated to AI to implement and oversee the national AI agenda. The establishment of such a mechanism would ensure policy coherence, define central standards and guidelines, monitor risks and drive further coordination across sectors and levels of government (see Box 6.4 for an international example). Refining and strengthening the roles and responsibilities of the CDTOs would also help institutionalise accountability and implementation capacity within the public sector.
In September 2024, Korea established the National Artificial Intelligence Committee, a high-level advisory body designed to steer the country’s AI strategy. The Committee is co-chaired by the President and a prominent private sector leader, embodying a collaborative approach between government and industry. The Committee is comprised of seven ministers from ministries, including the Interior and Safety, Science and ICT, Economy and Finance, Trade, Industry and Energy, Education, and Foreign Affairs, as well as five deputy-minister-level government officials. In addition, the industry and academia are represented by 29 experts. The Committee also has subcommittees focusing on specific AI domains such as ethics and safety, talent development and legal and institutional challenges.
The Committee is supported by the Secretariat, composed of four teams – Policy and System, Safety and Trust, Industrial Technology, and Talent Infrastructure.
The Committee’s primary responsibilities include:
Reviewing and coordinating national AI policies to ensure alignment across various governmental and private sector initiatives.
Developing strategies for research, development, and investment to identify priorities and allocate resources effectively.
Expanding AI infrastructure to facilitate the growth of data centres and computing resources.
Identifying and improving regulations to address legal and ethical considerations related to AI deployment.
Disseminating AI ethics principles to promote responsible AI practices across sectors.
Responding to sectoral changes, including adapting to AI-driven transformations in education, labour, and the economy.
Source: (Korean Government, 2024[14])
Institutional development needs to be supported by sustained investment. Ukraine has demonstrated agility in mobilising international support and forging public-private partnerships. However, the long-term success after the war requires a stronger strategic approach to digital government investments, including AI initiatives in the public sector. This would enable Ukraine to manage digital government investments across the full lifecycle, from planning to evaluation, in a coordinated way, aligning them with the national strategies and long-term priorities. Ukraine would also be able to target funding for essential, shared services and systems across the public sector (see more in Figure 6.2 and Box 6.5) (OECD, 2025[15]).
The figure below shows the pillars and dimensions of the framework.
Source: (OECD, 2025[15])
Australia has adopted a comprehensive, whole-of-government approach to coordinating digital and ICT investments, anchored in the mandate of the Digital Transformation Agency (DTA). The DTA operates as the central authority for guiding digital transformation across the federal public sector. Established to provide leadership on government service delivery, the agency is legislatively empowered to lead policy, provide strategic oversight, and ensure value-for-money outcomes in the planning, procurement, and implementation of digital initiatives. This integrated mandate reflects Australia's recognition that managing digital transformation, including the use of AI, requires not only technical capacity but also strategic governance throughout the full investment lifecycle.
Responsibilities of the DTA
The DTA's core responsibilities include developing and monitoring national digital strategies, setting ICT standards and policies, and overseeing digital and ICT investments across government. The DTA also advises the responsible Minister on digital investment proposals and ensures alignment with broader public sector reform goals. In addition, the agency manages a centralised procurement function, simplifying purchasing for government agencies while promoting the reuse of digital platforms and services.
Implementation through the Digital and ICT Investment Oversight Framework (IOF)
To operationalise its oversight function, the DTA introduced the Digital and ICT Investment Oversight Framework (IOF), a structured, end-to-end model for managing public sector digital and ICT investments across their lifecycle. The IOF is built around six interdependent stages that mirror the annual budget process:
Strategic Planning (Pre-Budget): Engages agencies in long-term investment planning to identify gaps, avoid duplication, and align proposals with whole-of-government digital goals.
Prioritisation (Pre-Budget): Assesses and ranks investment proposals based on six criteria, including alignment with government strategy, risk profile, and expected public value.
Contestability (Budget Phase): Provides a mechanism for reviewing and refining proposals through inter-agency consultation, ensuring adherence to policy, design standards, and technical feasibility.
Assurance (Implementation Phase): Implements oversight and risk management measures during delivery, tailored to the size and criticality of each investment.
Sourcing (Implementation Phase): Applies dynamic purchasing and procurement frameworks to ensure value-for-money and reuse of proven tools and services across the public sector.
Operations / Real-World Use (Ongoing): Focuses on data collection, performance tracking, and continuous monitoring of investment outcomes to inform future decisions.
In addition to considerations of governance and investments, data quality and digital infrastructure will be pivotal to scaling up the AI use in the public sector. Ukraine should build on its existing data exchange frameworks, such as Trembita, by improving data standards, interoperability and ensuring real-time access across registries and systems. Concurrently, increasing use of AI-driven services will require robust digital infrastructure. Ukraine should consider continued investment in a set of shared, secure and interoperable digital systems, designed to support broad access to public and private services, also known as digital public infrastructure (DPI) (see more in Figure 6.3 and Box 6.6). This would not only generate efficiency gains, but also promote resilience, security and scalability for the public sector to provide better and responsive services (OECD, 2024[18]). In addition, investing in DPI would contribute to building national capabilities to produce AI using its own data, infrastructure and workforce (OECD, 2025[1]).
The figure below shows the connection between base registries, digital public infrastructure, and services.
Source: (OECD, 2024[18])
Korea’s shared data centres and government cloud
Korea’s National Information Resources Service (NIRS) has been working with the Ministry of the Interior and Safety to upgrade key hardware, networks, and management tools to help modernise Korea’s technology and enable a migration to the cloud. A critical part of this has been the construction of new government data centres, which ensure compliance with government requirements, cost-efficiencies with a reduced technology footprint, and job creation and local investment in target areas. These data centres have also been made available to the Government’s main partners in the private sector, which helps to ensure that companies holding or handling sensitive data are doing so in an environment that meets the government’s requirements for security, back-up, and redundancy, among others. With measures around environmental considerations and renewable energy, the data centres help reduce the environmental impact of Korea’s digital government, particularly as it prepares to make greater use of AI solutions.
AI foundation models
Foundation models are advanced generative AI systems trained on extensive datasets, typically through self-supervision at scale, enabling their application across a broad set of tasks. Governments may either develop dedicated foundation models or customise the use of existing proprietary or open source models to fit national or public-sector contexts. Such customisation is achieved by fine-tuning the model, including additional training using targeted datasets, to improve its effectiveness in performing specific tasks or addressing particular issues.
Increasingly, governments are interested in developing national or regional foundation models as a means to strengthen technological sovereignty and accurately represent linguistic and cultural nuances. The most notable examples include:
Latam-GPT, developed collaboratively by over 30 institutions in Latin America to create an AI model trained specifically on regional data
OpenEuroLLM, an EU-funded project building open-source models to encompass all official European languages
Italy’s Minerva project, the first large language model fully developed and trained specifically for the Italian language built entirely from scratch.
Finally, it is crucial that the trustworthy use of AI remains a key guiding goal in all AI development and deployment to maintain public trust. As outlined in the White Paper on AI Regulation, ethical guardrails need to be operationalised through practical implementations (Government of Ukraine, 2024[7]). Each public sector AI initiative should be accompanied by an ex-ante impact assessment to identify potential risk and mitigation strategies from the outset, as well as a risk-based approach for managing potential risks that may arise. Ukraine could consider implementing initiatives promoting algorithmic transparency, especially for high-risk AI systems. Engaging civil society, academia, and service users in these processes would further reinforce legitimacy and help ensure that AI-driven services reflect the broad set of needs and rights of the population. These instruments are discussed in the following section on guardrails, including international examples.
What guardrails are in place for the use of AI in government?
Guardrails are policy tools that help governments set boundaries for the responsible, trustworthy, and human-centred use of AI in government and in particular to improve public service delivery (OECD, 2024[2]). These can include soft laws and guidance as standards, policies, laws and regulations, enforcement mechanisms, monitoring bodies (within the executive), oversight bodies (beyond the executive), or collective agreements with the public sector workforce (OECD, 2025[1]). Together with the enablers, these help to overcome some of the key risks and challenges, such as lack of accountability and oversight, insufficient transparency and explainability, risks arising from skewed data, and weak cybersecurity robustness, among others (OECD, 2025[1]). This analysis explores key guardrails for AI in government, outlining Ukraine’s current progress and identifying remaining gaps.
Progress to date
Ukraine has made significant progress in laying the groundwork for the trustworthy use of AI in government and, more specifically, in public service delivery. However, many initiatives remain at a preliminary stage. As shown in Table 6.2, the adoption of the Strategy for the Digital Development of Innovations in Ukraine for the period until 2030 (2024) (see more in section What enablers are in place for the use of AI in government?) and the launch of the AI Regulation Roadmap (2023) and White Paper (2024) (see more in Box 6.7) reflect a commitment to aligning with international standards, such as the EU AI Act (see more in Box 6.1). The strategy and roadmap have aimed to promote the adoption of AI in government by introducing measures, including a phased approach with voluntary tools, sectoral guidelines, and institutional capacity-building. These measures are gradually paving the way for more binding instruments, such as dedicated AI legislation and the establishment of a regulatory authority (Government of Ukraine, 2024[22]; OECD, 2025[23]). Ukraine has also introduced soft laws to guide AI development, including the AI Regulation White Paper (see more in Box 6.10), voluntary codes of conduct, and guidelines for public servants and other sectors (Government of Ukraine, 2023[24]; OECD, 2025[23]).
At the same time, the Government of Ukraine is preparing a draft Law on AI, which is expected to be in alignment with the EU AI Act (OECD, 2025[23]). Ongoing work on supporting institutional frameworks includes an upcoming National AI Development Strategy 2030, and a National Digital Agenda, which aim to position the country among the global leaders in the use of AI across society and public administration (OECD, 2025[23]) (see more in Table 6.2). These efforts are expected to prioritise ethical considerations, transparency, and human rights, as reflected by Ukraine’s adherence to the OECD Recommendation on Artificial Intelligence [OECD/LEGAL/0449], its signing of the Council of Europe’s Framework Convention on Artificial Intelligence (Council of Europe, 2024[25]), and other international commitments promoting the responsible and ethical use of AI (see more in Annex 6.A).
The country is building its regulatory and institutional capacity to support a future unified digital regulatory authority dedicated to overseeing AI development and ensuring compliance with forthcoming binding AI legislation. This aims to align the use of AI in government and society with the EU AI Act. Table 6.2 also presents that Ukraine is progressively developing AI technical guidelines, modernising its data protection law to align with the General Data Protection Regulation (GDPR), and launching initiatives such as the Responsible AI Centre to inform stakeholders about incoming legislation (Government of Ukraine, 2024[22]), and Glossary of Terms on AI (Government of Ukraine, 2024[26]). New methodologies, such as HUDERIA (see Table 6.2), are supporting risk assessment, transparency, and efforts to mitigate skewed outcomes (Government of Ukraine, 2024[7]). Additionally, cybersecurity institutions are incorporating AI into training programmes, and innovation procurement practices are beginning to evolve, though AI-specific clauses on ethics, transparency, or risk management are not yet included in public tenders (Government of Ukraine, 2024[7]). While formal regulatory mechanisms, such as audits, compliance inspections, and centralised oversight, are still evolving, Ukraine’s phased approach, prioritising soft regulation, is supporting its transition toward the development of a more robust AI governance and further alignment with the EU AI Act (OECD, 2025[23]; European Union, 2024[27]).
The table below assesses the guardrails that can be in place to strengthen the use of AI in government
Where to focus next
While Ukraine has made progress in using AI in government, substantial gaps remain in monitoring and oversight, and risk management. The MDT and its Expert Committee currently provide strategic oversight for AI initiatives, but Ukraine has not yet established a dedicated AI supervisory body or a central registry of government AI systems (Government of Ukraine, 2024[7]). This limits the country’s capacity to ensure consistent accountability and enforce standards on a large scale (OECD, 2025[1]). As AI deployment in public services expands, Ukraine is prioritising and working on the creation of a dedicated oversight and regulatory body with enforcement powers, as envisioned in its AI Regulation Roadmap (2023) (Government of Ukraine, 2024[22]). Improving central monitoring, such as through the recurrent mapping of AI initiatives currently conducted by the MDT, would help Ukraine prepare for future binding legislation aligned with the EU AI Act. This approach could support the consistent identification of risks and promote greater transparency and accountability in the use of AI and algorithmic systems (European Union, 2024[27]; OECD/UNESCO, 2024[40]).
Another priority area is the development of binding cross-sectoral guidelines for the design, development, and management of AI systems in public service delivery. While Ukraine has introduced initial soft law instruments, such as the guidelines for public servants and voluntary codes of conduct, there are still no legally binding operational standards required for public agencies to manage high-risk AI systems consistently (Government of Ukraine, 2024[7]; OECD, 2025[23]) This gap may hinder responsible deployment, especially in high-impact sectors (OECD, 2025[1]). Similarly, although innovation procurement is emerging, AI solutions have not yet been tailored to address the specific risks and requirements of AI technologies (OECD, 2025[23]) (see more in What enablers are in place for the use of AI in government?). Ukraine could update procurement frameworks to include AI-specific safeguards and integrate international best practices into upcoming GovTech initiatives to ensure that third-party AI systems align with the public sector’s values of fairness, transparency, and trust (OECD, 2025[1]).
Ukraine’s Roadmap for AI Regulation outlines a phased and bottom-up strategy aimed at ensuring the trustworthy use of AI (Government of Ukraine, 2023[24]). Developed by the Expert Committee on the Development of Artificial Intelligence in collaboration with the Ministry of Digital Transformation, the roadmap was developed through multi-stakeholder consultations that involved representatives from business, academia, civil society, and policymakers. Drawing from global best practices and anticipating the adoption of the EU AI Act, the roadmap aims to position Ukraine to align its AI governance with international norms while protecting the rights of citizens and promoting innovation (Government of Ukraine, 2023[24]). An emphasis is placed on the responsible integration of AI in public services, ensuring that solutions are safe, fair, and transparent before they are scaled across government.
The roadmap is structured into two phases (Government of Ukraine, 2024[22]; Government of Ukraine, 2023[24]):
Phase I (2023–2026) focuses on preparatory measures, including:
Development of a White Paper on AI Regulation.
Soft laws and introduction of specific guidelines, including for civil servants.
Introduction of voluntary codes of conduct.
Launch of a sandbox for safe AI and Blockchain experimentation.
Participation in the HUDERIA project to assess the human rights risks of AI.
Development of the Responsible AI Centre to inform regulatory updates.
Implementation of the Trusted Flaggers initiative to identify AI human rights violations.
Update of Ukraine’s National AI strategy by 2030.
Phase II (from 2026 onward) will involve the gradual adoption of legally binding AI legislation aligned with the EU AI Act, as well as the establishment of a dedicated AI regulatory body.
The roadmap prioritises giving businesses and public organisations the tools to self-regulate and prepare for future compliance based on principles of service delivery, partnership, and global alignment. Key principles include (Committee on AI in Ukraine, 2023[41]):
Service function: supporting innovation and public value before enforcement.
Balance: learning from global regulatory experiments before implementing domestic rules.
Self-regulation: co-developing norms with industry and public service institutions.
Global perspective: ensuring Ukrainian AI products and services are export-ready.
Product approach: focusing on practical tools for market access and compliance.
The OECD AI Principles provide a framework for the development and deployment of trustworthy AI that respects human rights and democratic values. These principles, first adopted in 2019 and updated in 2024, serve as an intergovernmental standard to guide policymakers and AI actors in shaping AI policies that maximise benefits while mitigating risks. They emphasise international co-operation and interoperability, ensuring AI systems contribute to economic growth, social welfare, and environmental considerations. Currently, the European Union, the Council of Europe, the United States, the United Nations, and other jurisdictions incorporate the OECD’s definition of an AI system and its lifecycle into their legislative and regulatory frameworks. These principles, along with the definition and lifecycle, are integral components of the OECD Recommendation on Artificial Intelligence.
Value-based principles
The OECD AI Principles promote AI that is innovative, trustworthy, and human-centric. They include:
Growth and well-being that benefits all groups in society: Stakeholders should proactively engage in responsible stewardship of trustworthy AI in pursuit of beneficial outcomes for people and the planet.
Human rights and democratic values, including fairness and privacy: AI actors should respect rule of law, human rights, democratic and human-centred values throughout the AI system lifecycle.
Transparency and explainability: AI Actors should commit to transparency and responsible disclosure regarding AI systems.
Robustness, security, and safety: AI systems should be robust, secure and safe throughout their entire lifecycle.
Accountability: AI actors should be accountable for the proper functioning of AI systems and for the respect of the above principles.
Recommendations for policymakers
To ensure AI is trustworthy and beneficial, policymakers are encouraged to:
Investing in AI research and development: Governments should support long-term public and encourage private investment in interdisciplinary R&D to drive trustworthy AI innovation.
Fostering a responsive AI-enabling ecosystem: Governments should develop dynamic, durable, and interoperable digital ecosystems for trustworthy AI.
Shaping an enabling interoperable governance and policy environment for AI: Governments should create agile policies that support the shift from AI research to real-world deployment.
Building human capacity and preparing for labour market transitions: Governments should equip people with skills to use AI effectively and adapt to changes in the world of work and society.
International co-operation for trustworthy AI: Governments and stakeholders should collaborate globally to uphold and advance trustworthy AI principles.
Source: (OECD, 2024[42])
Transparency around government use of AI is critical to maintaining democratic accountability, building public trust and enhancing internal knowledge-sharing around the technology. However, globally, many public sector institutions lack sufficient understanding and visibility into their own AI use, significantly hindering these objectives. Effective transparency measures enable governments to proactively inform citizens and stakeholders about AI systems without being prompted by requests.
Several countries have implemented proactive transparency measures, including public registries of AI systems. These registries serve as centralised, publicly accessible repositories detailing AI systems deployed by government institutions, their purpose, relevant sectors and jurisdictional coverage. Some of examples include:
Colombia’s dataset on automated decision systems in the Colombian public administration,
The UK’s Algorithmic Transparency Records,
National government public algorithm inventories in Chile, France and the Netherlands Public Algorithms Inventory, and
Sub-national algorithm registers Amsterdam, the Netherlands and Helsinki, Finland.
Publishing algorithm source code and documentation has similarly emerged as a measure of transparency. The practice of open sourcing the code for public algorithms is widely regarded as a best practice in algorithmic transparency. This is particularly beneficial for technical and expert audiences, as it enables them to examine, test and verify how these systems operate, which could in turn promote accountability and trust. Some efforts are still short of releasing the full source code but require public sector institutions to publish detailed documentation that can have a similar effect.
In France, the Digital Republic law mandates government agencies to “make publicly available, in an open and easily re-usable format, the rules defining the main algorithmic processing used in the accomplishment of their mission when such processing is the basis of individual decisions”.
The UK’s the Algorithmic Transparency Recording Standard (ATRS) mandates public sector organisations to transparently disclose details about their use of algorithmic methods in decision-making processes.
In Canada, the Directive on Automated Decision-making explicitly outlines, in detail, explainability requirements for AI systems differentiated by levels of risk, determined by use of an Algorithmic Impact Assessment tool (the results of which must also be published). As of 2025, the tool has matured into a robust, risk-aware system, enhancing systemic oversight and promoting responsible use across the public sector.
Source: (OECD, 2025[1])
In June 2024, Ukraine's Ministry of Digital Transformation released the White Paper on Artificial Intelligence Regulation in Ukraine, outlining a strategic framework for AI governance, building upon the previously published AI regulation roadmap (2023) (Government of Ukraine, 2023[24]). This document emphasises a phased and bottom-up approach to regulation, fostering self-regulation and innovation while preparing for alignment with the European Union's AI Act (Government of Ukraine, 2024[7]). The strategic goals and key initiatives for service delivery outlined in the document are as follows:
Strategic goals:
Enhancing AI industry competitiveness: Equip Ukrainian businesses to compete globally and integrate into international markets.
Safeguarding citizens' rights: Protect individuals from potential risks and misuse related to AI.
European integration: Harmonise Ukraine's AI regulations with EU standards to facilitate seamless integration.
Key tools and initiatives for service delivery:
Sandbox: Establishing a controlled environment where AI technologies can be tested within public services under regulatory oversight enables the assessment of AI applications' impact on service delivery and user experience before broader implementation.
AI system evaluation methodology: Develop a risk-based assessment framework to categorise AI systems by their impact on human rights and society, including voluntary AI labelling for public institutions. This initiative aims to inform citizens about the involvement of AI, fostering trust and understanding.
Sector-specific guidelines: Development of AI design and management guidelines for integrating AI into various policy sectors, ensuring that AI applications align with ethical standards and effectively meet public needs.
Monitoring and oversight
Centralised bodies or tools can support monitoring and oversight of the ethical, transparent, and effective use of AI technologies across government. Oversight bodies in the non-executive branch of government can support ensuring the consistent application of AI regulations and policies by ensuring compliance with legal and ethical standards through ex-post evaluations and audits (OECD/UNESCO, 2024[40]). Oversight bodies in the executive function can support managing risks, addressing ethical concerns, and streamlining the adoption of AI by monitoring AI projects at a large scale (OECD, 2024[2]).
Ukraine has not yet established a non-executive body to oversee AI solutions, although it is developing a concept for a unified digital regulator as envisioned in the country’s AI roadmap (Government of Ukraine, 2024[22]) (see more in Box 6.11). Executive supervision is currently carried out by the Cabinet of Ministers, the MDT and its Expert Committee, which coordinates AI initiatives but lacks monitoring authority (Government of Ukraine, 2024[7]). The Expert Committee provides strategic guidance, and monitors select government AI projects (OECD, 2023[43]; IFES, 2024[28]) (see more in Table 6.2). EU member states have adopted varied models for overseeing the use of AI in the public sector, combining institutional independence with expert involvement to uphold ethical and legal standards (OECD, 2025[44]). In France, multiple authorities have been tasked with protecting fundamental rights in line with the EU AI Act, ensuring AI systems align with ethical and regulatory standards (OECD, 2025[44]). Hungary has created the AI Coalition, led by the Ministry of Innovation and Technology, which brings together public institutions, academia, and industry to guide AI deployment (OECD, 2025[44]). Ireland follows a decentralised model, delegating enforcement responsibilities to sector-specific regulators while assigning overarching supervisory roles to designated authorities (OECD, 2025[44]). Though tailored to each country’s governance system, these approaches share a common goal: fostering responsible AI innovation while ensuring transparency, safety, and accountability.
To align with the EU AI Act and national objectives outlined in the AI Regulation White Paper, Ukraine is advancing plans to establish an independent regulatory body for the digital sector. Ukraine is currently developing a concept for a unified digital regulator, which will include financial estimates and technical specifications.
The proposed institution, the National Commission for the State Regulation of Digital Spheres (provisional name), is expected to function as an independent central executive body with special status, ensuring autonomy in decision-making. Its mandate would span a broad range of areas requiring regulatory oversight, including artificial intelligence, digital services and markets, data governance, public information access, and online child protection.
This unified authority is envisioned as a practical response to the complexity of implementing EU digital regulations, offering an integrated solution that avoids the inefficiencies of creating multiple specialised agencies. It aims to meet EU standards for regulatory bodies, including independence, transparency, impartiality, and adequate financial, technical, and human resources.
The Draft Law on the Digital Regulator is expected to be adopted by the end of 2026. To support its effective establishment and alignment with the EU AI Act, Ukraine will benefit from closer collaboration with EU institutions, participation in the European Artificial Intelligence Board, and access to technical assistance programmes to support implementation of best practices in AI oversight. Building institutional capacity will be essential to ensure effective regulation across interconnected digital domains as Ukraine advances toward EU membership.
Tools for monitoring, transparency standards, and skewed data detection can also support effective oversight of AI use in the public sector. Monitoring tools to assess the effectiveness and phased deployment of AI investments can help to align AI applications with objectives and identify opportunities for scaling across government (OECD/UNESCO, 2024[40]). Countries are deploying tools to ensure that AI systems do not amplify issues arising from skewed data or introduce new safety risks (OECD/UNESCO, 2024[40]). These measures enhance algorithmic transparency, which is crucial for effective AI oversight. Additionally, AI impact assessment frameworks can save human, financial, and environmental resources, boost service quality and trust, and ensure relevance and context (OECD/UNESCO, 2024[40]; OECD, 2025[1]). Ex-ante and ex-post impact assessments help predict consequences, mitigate risks, prevent harm, engage citizens, and tackle societal challenges (OECD, 2024[2]; OECD, 2025[1]).
Ukraine has begun developing tools related to monitoring, transparency, and identifying issues linked to skewed data in AI systems, but these remain at a nascent and mostly non-binding stage. The country has not yet established a dedicated executive function to monitor AI solutions across the public sector; however, the MDT is currently carrying out AI mapping exercises to track and coordinate AI use within government (Government of Ukraine, 2024[22]). However, current strategic documents and guidelines for public servants endorse transparency as a guiding principle, and there are intentions to adopt algorithmic transparency tools and standards (Government of Ukraine, 2024[7]). Moreover, the country has acknowledged the challenges associated with skewed data in AI systems and has embedded principles aimed at preventing adverse outcomes in its guidance documents (Government of Ukraine, 2024[7]) (see more in Table 6.2). Ukraine can learn from the UK's Model for Responsible Innovation which supports public sector teams in responsibly developing and deploying data-driven technologies and AI (Box 6.12).
The Model for Responsible Innovation, developed by the UK’s Responsible Technology Adoption Unit (RTA) within the Department for Science, Innovation and Technology, is a hands-on tool to help public sector teams innovate responsibly with data and AI. It was developed through testing and consultation across various government departments and is now publicly available to support the adoption of ethical AI, aligning with the OECD principles for trustworthy AI.
The RTA team utilises the model to conduct free red-teaming workshops, enabling public sector teams to rapidly identify and address risk issues in their AI use cases. These workshops have already been applied in policy sectors such as justice, health, social care, and energy, helping teams refine their tools before deployment. The model aims to establish justified trust in AI systems, particularly those that have a significant impact on the public. It achieves this by guiding teams through eight core fundamentals of responsible innovation:
Transparency: Open systems and explainable decisions
Accountability: Clear governance and responsibility
Human-centred value: Systems designed to benefit people
Fairness: Equitable outcomes and processes
Privacy: Protection of personal data
Safety and security: reliable, robust, and non-harmful tools
Societal well-being: systems that support the public good
These are supported by six conditions: from meaningful engagement to robust technical design and effective governance, that must be met to ensure public services powered by AI remain fair, ethical, and fit for purpose.
The model is instrumental in service-delivery purposes, helping teams working on chatbots, automated eligibility systems, or data-linking platforms to identify ethical risks early and apply mitigation strategies. Tools such as the DESNZ Project Delivery Chatbot and the Better Outcomes through Linked Data (BOLD) programme have benefited directly from this approach.
Source: (UK Government, 2024[46])
Guidelines on the design, development, and management of AI in government
AI guidelines for public sector organisations provide a structured framework to ensure that AI applications are implemented responsibly, ethically, and effectively (OECD/UNESCO, 2024[40]). These guidelines are crucial as they help maintain public trust, ensure compliance with legal and ethical standards, and promote transparency and accountability in AI systems (European Commission, 2024[47]). By offering clear guidance on best practices, risk management, and ethical considerations, such guidelines facilitate the adoption of AI for public service delivery by enhancing efficiency, improving decision-making, and ensuring that AI systems are aligned with public needs (Government of Sweden, n.d.[48]; OECD/UNESCO, 2024[40]).
Ukraine’s non-binding AI guidelines for public servants represent a significant first step in building public sector AI governance. Developed as part of a broader White Paper on AI regulation (Government of Ukraine, 2024[7]), they aim to equip public servants with practical and ethical principles for responsible AI use. The document emphasises human rights, data protection, and ethical conduct (Government of Ukraine, 2025[34]). It offers practical advice on prompt creation, the use of generative AI tools, and AI literacy, along with guidance on cybersecurity hygiene and risk management, while referencing international standards (Government of Ukraine, 2025[34]).
Ukraine can learn from Sweden's approach to improving AI adoption through public sector-specific guidelines. Developed by the Swedish Agency for Digital Government (Digg) and the Swedish Authority for Privacy Protection (IMY), these guidelines provide a comprehensive framework to ensure the responsible and effective use of AI technologies (Government of Sweden, 2025[49]). Within 18 guidelines, the framework covers 7 dimensions: management and responsibility, data protection and personal data processing, labour law, procurement, information security, copyright, and ethics (Government of Sweden, n.d.[48]).
The management and responsibility dimension emphasises the need for clear accountability and governance structures to oversee AI projects. Data protection and processing compliance are crucial to ensure that AI systems respect individuals' privacy rights. Labour law considerations address the impact of AI on the workforce, ensuring fair treatment and upskilling opportunities. Procurement guidelines ensure that AI solutions are acquired transparently and ethically, while information security measures protect AI systems from cyber threats. Copyright guidelines address the use of AI-generated content, and ethical considerations ensure that AI applications align with societal values and human rights (Government of Sweden, n.d.[48]).
To support the adoption of these guidelines, Sweden has developed complementary materials and capacity-building initiatives (Government of Sweden, 2025[49]). These include webinars aimed at increasing the competence of public sector employees in handling AI technologies. Additionally, the guidelines are continuously updated to reflect the latest developments in AI technology and legislation. The establishment of a dedicated website serves as a central hub for all information related to AI in the public sector, providing access to guidelines, best practices, and other resources (Government of Sweden, n.d.[48]).
While these guidelines are a critical entry point, they remain largely individual-focused and voluntary in nature. To strengthen public trust, ensure safety, and enable the scalable deployment of AI in government, Ukraine should transition from individual public servant-level guidance to a comprehensive organisational-level approach, including capacity-building support. Ukraine can learn from international guidelines such as the G7 Toolkit for Artificial Intelligence in the Public Sector. This toolkit offers practical insights and good practices for AI use, integrating ethical considerations and showcasing public sector AI use cases (OECD/UNESCO, 2024[40]). Similarly, Governments such as Sweden, the UK, and the US have developed national organisational guidelines and playbooks to ensure compliance with AI standards (Box 6.13).
Institutionalising experimentation practices is equally important. Beyond providing guidance, Ukraine could also embed structured AI experimentation spaces and practices across the public administration. Spaces such as sandboxes, labs, and innovation units support the testing and diffusion of these practices, enabling the safe and iterative experimentation of AI systems, the co-design of solutions with users, and the systematic collection of evidence to inform decisions on scaling up innovations (see Box 6.14). However, institutionalising such practices through organisational-level guidelines also helps ensure that experimentation is not ad hoc or confined to specific spaces but becomes a sustained capacity within innovation processes across government, while also creating permission to take balanced risks.
Innovation teams within public administrations are increasingly adopting AI across the innovation cycle, from research and ideation to prototyping, implementation and evaluation. The OECD is working on a forthcoming publication that will present a collection of practices, including evidence from countries such as Austria, Chile, Korea, the UK, Singapore, Ukraine, and others. Current evidence shows that AI can complement innovation teams in five distinct roles with various practices:
Insight finder: Accesses, aggregates, and analyses large datasets to frame problems, benchmark performance, and design evaluation strategies.
Pattern detector: Interprets structured and unstructured data to reveal correlations, outliers, and trends that inform decision-making.
Idea booster: Supports creative and strategic thinking by proposing innovative concepts, framing trade-offs, and offering alternative models.
Scenario tester: Simulates outcomes of policy options or prototypes, stress-testing designs and reducing uncertainty.
Message shaper: Tailors and delivers information clearly through stakeholder mapping, visualisation, and persuasive communication.
AI roles are not confined to one stage. For example, insight finding and pattern detection are central to research and evaluation, while idea boosting and scenario testing are critical during ideation and prototyping. Message shaping is a cross-cutting role that strengthens communication at all stages. However, for AI to deliver its full value across innovation processes and their outcomes, governments also need to address enabling conditions such as:
Data and technology availability, including access to reliable, interoperable, and proprietary datasets.
New competencies, including AI literacy and reinforced innovation skills across the public workforce.
Ethics and accountability frameworks to safeguard transparency, fairness, and trust in AI-enabled processes.
Evaluation mechanisms to measure the effectiveness and long-term impacts of innovations, including AI-powered solutions.
Source: (OECD, 2026, forthcoming[55])
Procurement
Procurement processes for third-party AI solutions in the public sector involve the systematic acquisition of AI applications from external vendors (Monteiro, Hlacs and Boéchat, 2024[50]); (OECD, 2024[51]). These processes are crucial as they ensure that AI solutions are selected based on clear criteria, including ethical standards, transparency, and compliance with legal requirements (Monteiro, Hlacs and Boéchat, 2024[50]). Moreover, innovative procurement mechanisms such as GovTech programmes can support the adoption of AI in government by streamlining the participation of startups and cutting-edge SMEs, problem-driven procurement, facilitating experimentation and piloting, embedding ethical and performance standards early, and promoting collaborations within innovation ecosystems (OECD, 2025[1]; OECD, 2024[51]).
Although innovation procurement is emerging in Ukraine, current procurement rules have yet to address the specific challenges of AI systems. While the AI roadmap highlights the need to manage public-private partnerships in AI securely, current tender processes lack specific clauses or standards addressing AI-related ethics, performance, or risk management (OECD, 2025[23]; Government of Ukraine, 2024[22]) (see more in What enablers are in place for the use of AI in government?). However, the country is beginning to explore innovative procurement approaches to increase the adoption of AI in government. As part of its Digital Innovation Development Strategy until 2030 – WINWIN (2024), Ukraine aims to establish a variety of different types of GovTech initiatives, including a dedicated GovTech programme to support the procurement of third-party innovative solutions (Government of Singapore, 2025[52]). OECD countries are increasingly developing targeted guidelines to support responsible and innovative AI procurement. For instance, the UK government has introduced policies and guidelines to manage AI procurement, ensuring that AI systems are safe, minimise risks related to skewed data, and respect privacy (UK Government, 2020[53]). Similarly, Singapore's National Artificial Intelligence Office has developed AI procurement frameworks to ensure the responsible use of AI in public services, and it collaborates on research to create solutions that are not yet commercially available (Government of Singapore, n.d.[54]) (Box 6.15).
Ukraine can learn from Singapore's approach to a structured procurement process and established research collaboration programme. The Singapore Public Sector AI Playbook provides a robust guide for government agencies to adopt and implement AI technologies effectively through procurement and applied research.
The Procure from the Industry mechanism outlines the Green Lane Procurement approach, which allows agencies to procure technology products and services from firms qualified under GovTech-endorsed programmes. This approach aims to improve process efficiency during tender and quotation evaluations, and support the growth and scaling of Singapore-based tech companies and startups. If these tender offerings are unsuitable, agencies can explore other sourcing methods.
The playbook also offers a mechanism for Research Collaboration to develop AI solutions for complex and unique problems. The initiative supports experimental projects through an established programme where an AI minimum viable model can be developed and deployed within 9 to 18 months, providing matching funds of up to $250,000 per project. The programme aims to create AI models that can be later integrated, deployed, and maintained by GovTech or in partnership with commercial solution providers. Research grants are also available for deployment and scalation of tested solutions.
How are various stakeholders being engaged to support the use of AI in government?
The engagement dimension considers who governments should engage with when developing and implementing the enablers and guardrails, as well as the development and use of AI systems. It includes the different stakeholders that should be engaged in building the foundations for a responsible use of AI in the public sector. In addition to the internal and external users of the AI systems, it also considers the various actors across (including other parts of the public sector, the private sector, civil society, academia, the public, etc.) to be engaged through targeted actions to effectively address policy challenges related to the use of AI in the public sector.
Progress to date
Ukraine’s notable progress on the use of AI in the public sector has been underpinned by active engagement with various stakeholders, including partnerships with the international community for AI. The government has recently established the WINWIN AI Centre of Excellence to serve as an international hub for AI integration in public services. The Centre could play an important role in bringing a variety of different types of actors together to develop, experiment and test AI services. For instance, the Centre plans to gather project proposals from both the public and private sectors and leverage internal and external expertise to pilot AI services to meet the public’s needs. In addition, the Expert Committee on the Development of AI in Ukraine (see Table 6.1) brings together experts from the public and private sectors, academia and civil society to increase Ukraine’s competitiveness in the area of AI.
Moreover, Ukraine’s advances in digital government and AI have also opened avenues for international partnerships. Many of Ukraine’s AI initiatives have been informed by the expertise and experience of peer countries, including the UK, Estonia, Sweden and Switzerland. Multilateral organisations such as the OECD and the UN have also played a supportive role in the development of Ukraine’s AI policies and practices, bringing a broad range of lessons learned and practices from the international community (OECD, 2025[23]). Furthermore, Ukraine has contributed to international discourse and standard setting around AI. For instance, Ukraine has contributed to the development of the Council of Europe’s Framework Convention on AI and the HUDERIA Methodology to assess the impact of AI systems on human rights (Government of Ukraine, 2023[24]), and is currently conducting a mapping of AI initiatives across the public sector with the support of the Council. Table 6.3 presents some of Ukraine’s key engagements around AI.
The table below assesses the engagement practices that should be in place to strengthen the use of AI in government.
Where to focus next
To sustain and expand stakeholder engagement at both national and international levels, Ukraine should consider systematically engaging them in all stages of AI initiatives and continue to leverage international expertise and support.
As Ukraine plans to deploy more AI-driven pilot projects across government, the government should consider conducting regular user testing and establishing feedback mechanisms throughout the development and use of these services. It is also important to engage civil society and experts in AI policy discussions and design. This will help ensure that AI-driven services benefit from a broad range of insights, are transparent and respect users’ rights.
The WINWIN AI Centre of Excellence (Box 6.3) presents a great opportunity to systematically cultivate partnerships with the private sector, civil society and academia to co-create AI-driven services. It would be important to monitor the Centre’s activities and report its progress and performance against defined objectives and goals transparently to the public. This would reinforce its legitimacy and secure continued support from both internal and external stakeholders. In addition, the Centre would be a great platform to provide a community of practice for the Ukrainian and international public sectors to share use cases and lessons learned (see international best practice on Box 6.16).
The UK has established a cross-government AI Community of Practice to support knowledge sharing, ethical awareness and collaboration around the use of AI in the public sector. This initiative responses to the growing need for public servants, regardless of their formal roles in AI, to better understand the implications, opportunities and challenges of AI in government.
The Community of Practice is open to all public sector employees with an interest in AI. Its purpose is to share information across departments, enabler peer-to-peer learning, and support the ethical and safe use of AI tools. The Community meets monthly online to discuss use cases, AI ethics and security.
Source: (Government of the UK, 2024[60])
Furthermore, deepening international co-operation through exchanges with partner governments and organisations remains essential, ranging from sharing best practices on using AI for public services to initiating joint pilot projects. Such cross-border engagements would not only support Ukraine’s rapid public-sector transformation but also provide valuable practices for the international community on effectively deploying AI under challenging conditions. By fostering ongoing collaboration and alignment with international partners, Ukraine can ensure its AI-driven transformation is durable, secure and interoperable on a global scale.
How is the impact of AI solutions on government being monitored?
Building on the pillars of enablers, guardrails, and engagement, the monitoring and evaluation of AI use in government should be strengthened to ensure it delivers positive outcomes (OECD, 2025[1]). While many governments are running AI pilots and experiments, there is limited monitoring and evaluation to determine which projects are effective and impactful (OECD, 2025[1]). This makes it difficult to demonstrate return on investment (ROI) and identify which initiatives should be scaled or replicated (OECD, 2025[1]). The following analysis focuses on how AI solutions are monitored in relation to key service delivery objectives, such as enhancing internal operations, designing seamless and accessible services, measuring and improving service delivery, and supporting experimentation in service delivery, in alignment with the OECD Recommendation of the Council on Human-Centred Public Administrative Services (OECD, 2024[61]).This focus reflects AI’s potential to boost quality and responsiveness in public service delivery. This section outlines Ukraine’s current progress and identifies key remaining challenges.
Progress to date
Ukraine has emerging practices for monitoring the use of AI in the government’s internal operations and service delivery design. As shown in Box 6.3, the WINWIN AI Centre of Excellence monitors experimental AI solutions in critical policy sectors. By 2025, the Centre had carried out 16 AI pilot projects, with half progressing toward minimum viable products (Government of Ukraine, 2025[62]). In service design, Ukraine has integrated AI into digital government platforms, such as Diia, prioritising seamless and accessible experiences. Emerging AI tools, including a virtual assistant for the Diia app, and an AI-powered consular spokesperson chatbot, are being evaluated for accuracy and responsiveness (UNN, 2024[63]; Government of Ukraine, 2024[64]). The government has also undertaken large-scale audits of digital accessibility, improving compliance with web accessibility standards (UNDP, 2024[65]). Within internal operations, AI-driven solutions developed by the MDT, such as the Digital Analysis of Legal Acts and AI analysis and translation of EU legislation, are assisting legal teams in analysing and monitoring regulatory documents (Government of Ukraine, 2025[66]).
AI practices are being developed to monitor service delivery performance and support experimentation. Table 6.4 explains that, while AI-specific performance tracking remains limited, Ukraine systematically measures user satisfaction and service efficiency. Diia’s data-driven approach helps refine digital services, and AI-assisted analytics are planned to enhance real-time feedback processing (UNDP, 2025[67]; OECD, 2025[23]). In social services and justice, AI pilots such as job-matching platforms for displaced persons and AI-assisted case sorting in courts are being tested for efficiency gains (Limantė and Moskvytyn, 2024[68]). Ukraine also fosters AI experimentation, ensuring that AI solutions undergo structured monitoring before scaling (Government of Ukraine, 2025[62]). The recently launched AI sandbox will provide support for adopting public-sector AI applications, allowing agencies to assess AI risks and ethical considerations (Government of Ukraine, 2023[24]). AI-driven services, such as the Mriia Educational App, which support students and parents in improving educational outcomes and reduce the administrative burden for educators, have successfully transitioned from pilot projects to consolidated solutions scaling to more than 2000 schools (Government of Ukraine, 2025[69]) (see more relevant initiatives in Box 6.17).
Ukraine’s efforts to integrate AI across the public sector, from administration to education and environmental protection, have aimed to enhance responsiveness, reduce manual workload, and align with EU integration and national strategic objectives. Several government initiatives illustrate this strategic shift toward AI‑embedded public services, focusing on enhancing internal operations, and designing seamless and accessible services. Relevant examples, among others, include:
Diia AI Assistant - Ministry of Digital Transformation: A conversational assistant that will be built into Diia’s web and mobile platforms, which will help citizens navigate services based on life events, such as guiding users through obtaining a specific certificate or finding the nearest Administrative Services Centre (Government of Ukraine, 2025[70]). In the future, it will support document submission, track application progress, and offer voice interactions, enhancing accessibility and reducing user burden (Government of Ukraine, 2025[70]).
Mriia Educational App and Web Portal - Ministry of Digital Transformation with Ministry of Education: Mriia is Ukraine’s national educational app and web portal, designed for students, parents, and teachers. It offers smart scheduling, homework tracking, secure communication, and a rich content library, powered by AI-driven recommendations that tailor learning materials to each student’s needs and interests (Government of Ukraine, 2025[69]). Launched in a pilot during September 2024 and now scaling to 2,000+ educational institutions, it aims to reduce the administrative burden for educators, automating tasks, such as timetable planning, and create a personalised, engaging learning experience for all users (Government of Ukraine, 2025[69]).
Diia.Education - Ministry of Digital Transformation: This government platform provides relevant knowledge and skills, aiming to enable Ukrainian citizens to learn new professions in innovative formats and at no cost, including a track on artificial intelligence literacy (Government of Ukraine, 2025[71]). The web platform provides personalised content recommendations and enhanced learning engagement using AI (East Europe Foundation, 2023[72]).
National Ukrainian-Language LLM - Ministry of Digital Transformation: Ukraine’s Ministry of Digital Transformation, through its WINWIN AI Centre of Excellence and in collaboration with Kyivstar, has initiated the development of a national Ukrainian-language large language model (LLM) to enhance AI capacity and affirm digital sovereignty (Government of Ukraine, 2025[73]). The model is being trained exclusively on open-source Ukrainian data, including dialects, institutional terminology, and cultural references, and hosted entirely within national infrastructure to meet requirements for secure processing in sensitive sectors (Government of Ukraine, 2025[73]). Targeted for release by December 2025, the LLM is intended to support a range of public and private sector applications: digital assistants integrated into platforms such as Diia, tools for legal and legislative review, and personalisation functions within national educational technology systems (Government of Ukraine, 2025[73]).
Natalka Diia AI Chatbot - Ministry of Digital Transformation: Natalka assists potential partners in integrating with the Diia platform. The chatbot answers routine questions and helps choose the appropriate integration path before escalating cases to human specialists when needed (The Odessa Journal, 2025[74]). In five months, it handled over 2,300 requests, reducing the workload of specialists by 30% (The Odessa Journal, 2025[74]).
HR Assistant for Job Vacancy Generation - Ministry of Digital Transformation: Generates professional, well-structured job postings automatically, with creative ability where appropriate, saving time and ensuring consistency across public sector hiring (The Odessa Journal, 2025[74]).
Digital Expertise of Legal Acts - Ministry of Digital Transformation: Currently under pilot, this AI system supports legal teams by rapidly analysing and uploading digital legal acts into central systems. Currently processes about 35% of such documents within 72 hours, improving review speed and consistency (The Odessa Journal, 2025[74]).
AI Analysis and Translation of EU Legislation - Ministry of Digital Transformation: This AI-powered system scans Ukrainian laws, compares them with EU regulations, and flags alignment gaps, accelerating legal harmonisation critical for Ukraine’s European integration (The Odessa Journal, 2025[74]).
Deep Green Ukraine - State Forestry Agency, Space Research Institute, and ForestCom NGO: Combines satellite imagery with open data to detect illegal logging in near real time. Piloted in several regions, the tool maps forest loss, checks licensing, and enables faster enforcement (Digital Security Lab Ukraine, 2024[35]).
Victoria, AI Consular Spokesperson - Ministry of Foreign Affairs: An AI-generated avatar delivers official consular updates via video, with human-authored scripts, ensuring accuracy (Government of Ukraine, 2024[64]). Accompanied by QR codes linking to verified text, Victoria increases transparency, frees diplomats for complex tasks, and ensures 24/7 public communication (Government of Ukraine, 2024[64]).
SkillLab, an AI Skill-to-Job Matching Platform - State Employment Service: In collaboration with the Ministry of Economy, the Ministry of Digital Transformation, the Norwegian Refugee Council, and the World Bank, this AI-driven system captures users’ skills and job experiences to match them with employment vacancies and training programmes (SkillLab, 2025[75]). Integrated with the Ukrainian State Employment Centre and Diia Single Sign-On, it has supported over 12,000 job seekers in finding new career paths (SkillLab, 2025[75]).
Source: OECD based on cited references.
The table below assesses the measures that should be in place to monitor the impact of AI solutions for service delivery.
Where to focus next
With ongoing investment in AI for measuring and experimenting with service delivery, Ukraine can enhance its AI-driven public services' efficiency, effectiveness, fairness, and accountability. While the country tracks some metrics, there is no standardised system for monitoring how AI affects their internal process results or service outcomes, such as error rates or efficiency savings, and evaluations often rely on project-specific KPIs (UNDP, 2025[76]; OECD, 2025[23]). Policy efforts acknowledge these gaps, aiming to introduce compliance standards and internal monitoring systems (Government of Ukraine, 2023[24]). Experimentation is emerging but often outpaces formal evaluation, with pilots such as chatbots or AI avatars being launched before publicly available impact assessments (OECD, 2025[23]). To ensure AI-driven services are trusted and effective at scale, Ukraine needs consistent risk-based evaluation frameworks, public reporting, and risk assessment, especially as it aligns with the EU AI Act (OECD, 2024[2]).
AI adoption for measuring and improving service delivery
Monitoring the impact of AI in public service delivery involves systematically evaluating how AI systems affect efficiency, effectiveness, and accessibility. This process relies on collecting and analysing data, user feedback, and performance indicators to assess the outcomes of AI-enabled services, both to identify what is working and to guide continuous improvement (OECD, 2024[2]). Effective monitoring ensures that AI tools enhance service quality, reduce administrative burdens, reach their intended users, and avoid unintended consequences such as issues arising from skewed data, gaps in access, or system failures (OECD/UNESCO, 2024[40]). By grounding AI adoption in evidence, governments can optimise and scale successful solutions, build public trust through transparency, detect potential risks early, and promote ongoing policy learning and innovation (OECD, 2024[2]).
Ukraine tracks key user-centric indicators such as service uptake, satisfaction, and cost savings, though AI-specific monitoring remains limited and ad hoc. Existing measurement practices, such as those employed on the Diia platform, provide a foundation for more accountable AI use with demonstrated efficiency gains (OECD, 2025[23]; EU4DigitalUA, 2021[77]) (see more in Table 6.4). Ukraine could expand its use of systematic monitoring tools, such as a dedicated dashboard to track AI performance for service delivery (OECD, 2025[1]). While most countries do not currently monitor AI performance, countries such as Singapore have developed a whole-of-government monitoring dashboard to track the performance of government websites and digital services, including those enabled by AI (OECD, 2025[1]). The dashboard tracks website traffic, user feedback, and performance benchmarks, generating automated reports with key metrics to improve user experience (Government of Singapore, n.d.[80]) (see more in Box 6.18).
Singapore's Whole-of-Government Application Analytics (WOGAA) is a centralised platform designed to monitor and enhance the performance of government websites and digital services.
WOGAA provides real-time tracking using Real User Monitoring (RUM) and Synthetic Monitoring tools, consolidating key metrics such as traffic, user feedback, and performance benchmarks into a single dashboard. This approach supports government agencies to assess their digital services, including those enabled by AI, identify areas for improvement, and streamline communication across public organisations. WOGAA enables agencies to make data-driven decisions that enhance user experience and service delivery by offering automated reporting on daily, weekly, and monthly metrics.
WOGAA utilises advanced analytics to evaluate how digital services and AI-powered tools engage with users, ensuring they meet predefined standards of accuracy, responsiveness, and accessibility. Additionally, WOGAA's Smart Sentiments Summary feature leverages AI to analyse user feedback in nearly real-time, helping agencies identify key issues to improve user experience.
AI adoption for experimenting with service delivery
Monitoring the impact of AI in experimentation in service delivery involves structured oversight of pilots or early-stage AI developments before they are scaled. AI-driven policies and services can be tested in controlled environments, such as regulatory sandboxes, AI incubators, or innovation labs, to evaluate their development, deployment, and performance. The goal is to understand what works, identify risks, and ensure AI solutions are scaled responsibly based on clear evidence (OECD, 2025[1]). Effective monitoring includes setting success criteria (such as effectiveness, balanced outcomes, and usability), tracking technical metrics (such as accuracy and error rates), collecting user feedback, and identifying unintended effects (such as issues linked to skewed data or gaps in access), and adjusting projects that fail to meet ethical, legal, or performance standards (OECD/UNESCO, 2024[40]; OECD, 2025[1]). This approach helps countries create safe spaces for innovation, manage risks, adapt regulations, and build public trust (OECD, 2024[2]).
Ukraine is emergingly monitoring AI experimentation initiatives for service delivery. Initiatives such as the WINWIN AI Centre incubate and monitor pilot projects to determine their readiness for scale-up (Government of Ukraine, 2025[62]). The country has also launched an AI and Blockchain sandbox, providing an environment for testing and evaluating AI systems to promote safe and compliant development (Government of Ukraine, 2023[24]). Moreover, several AI pilot projects, such as chatbots for emergency services (UNN, 2024[63]), the DeepGreen monitoring tool (Digital Security Lab Ukraine, 2024[35]), and AI-powered education apps (EU4Digital, 2024[79]), are being monitored through testing and user feedback (see more in Box 6.17 and Table 6.4). Ukraine could benefit from more structured experimentation frameworks within its current AI sandbox and incubator, allowing for the systematic testing, evaluation, and scaling of AI solutions for service delivery. Best practices from OECD countries include the UK’s Regulatory Sandbox, which helps organisations develop, monitor, and evaluate innovative, data-driven services while complying with data protection regulations (see more in Box 6.19) (UK Information Commissioner's Office, 2024[81]).
Ukraine could benefit from the UK's example by establishing a similar regulatory sandbox to support and monitor public and private experimentation with AI. The UK Regulatory Sandbox, developed by the Information Commissioner’s Office (ICO), is an initiative designed to support public and private organisations in creating innovative products and services that safely and responsibly utilise personal data (UK Information Commissioner's Office, 2024[81]). While monitoring experimentation of innovative solutions, this free service provides participants access to the ICO’s expertise, helping them navigate data protection regulations and embed data protection by design into their projects (UK Information Commissioner's Office, 2024[81]). By fostering a controlled environment for experimentation, the Sandbox allows organisations to test:
Innovations related to the UK Children’s Code,
Innovations related to data sharing (particularly within the health, central government, finance, higher and further education, or law enforcement sectors).
Products and services exploring the use of innovative technologies, such as privacy-enhancing technologies and distributed ledgers.
The service helps organisations to ensure compliance with the UK General Data Protection Regulation (UK GDPR) and the Data Protection Act 2018 (DPA 18) (UK Information Commissioner's Office, 2024[81]). This approach mitigates risks and promotes the development of AI solutions that are effective and ethically sound.
Source: OECD based on mentioned sources.
Ukraine has made important strides in establishing the foundations for the trustworthy use of AI in the public sector, though many efforts are still in the early stages. As implementation advances, this section presents a set of sequenced policy recommendations – first by time horizon (immediate priorities feasible under current wartime and institutional conditions; recovery and reconstruction actions as capacity expands; and longer-term development reforms to align with OECD standards). As a second lens, each recommendation is also classified under the OECD Governing with AI Framework’s four pillars: Enablers, Guardrails, Engagement, and Impact Monitoring. Building on Ukraine’s ongoing digital transformation, these recommendations aim to promote the responsible and effective adoption of AI across the public sector.
Immediate priorities
Operationalise the ethical guardrails included in the strategic documents (Enabler)
Engage civil society, academia and users to better understand their concerns around the use of AI in the public sector.
Develop concrete initiatives to ensure the ethical guardrails are implemented to prevent or mitigate concerns, such as Canada’s Algorithmic Impact Assessment tool (Box 6.9).
Strengthen Ukraine’s AI skills development across the public service (Enabler)
Ukraine has made important progress through NAUCS, the CDTO Campus, and Diia.Digital Education, but these initiatives could be expanded into a comprehensive AI skills framework.
Develop a whole-of-government AI skills framework to clarify the different levels of AI literacy needed across the workforce (general users, managers/leaders, data and business experts, AI specialists).
Expand AI training by scaling NAUCS programmes into continuous professional development and expanding the CDTO Campus curriculum to include strategic foresight, risk management, procurement, innovation and experimentation, and ethical oversight.
Mainstream AI literacy through Diia.Digital Education, offering practical modules on generative AI, data security, identifying issues linked to skewed data, and prompt design, with tailored tracks for frontline AI system users.
Strengthen talent pipelines and collaboration by creating multidisciplinary AI project teams, fostering communities of practice, and using flexible talent models (fellowships, secondments, GovTech partnerships) to attract and retain expertise.
Strengthen AI design and management (Guardrail)
The MDT could prioritise the development of practical guardrails by focusing on transitioning from individual public servant-level guidance to a comprehensive organisational-level approach to guide the design and management of AI systems in government, including capacity-building.
Develop AI guidelines for public sector organisations in partnership with experts in AI, law, ethics, and public administration, transitioning from individual public servant-level guidance to a comprehensive organisational-level approach, including associated capacity-building support.
Align guidelines with OECD best practices and international standards, such as the Council of Europe’s Framework Convention on AI, and the EU AI Act, covering governance, data protection, procurement, cybersecurity, copyright, and ethical use.
Conduct public consultations and implement the guidelines through training programmes and workshops.
Engage a broad variety of stakeholders in all stages of AI initiatives systematically (Engagement)
Ukraine should consider conducting user testing regularly and establishing feedback mechanisms throughout the development and use of AI-driven services.
Engage civil society, experts and the public in AI policy discussions and design to ensure responsive, transparent and ethical AI services.
Deepen international co-operation in the area of AI in the public sector (Engagement)
Continue aligning national policies and practices with the EU AI Act, including risk-based classifications, transparency requirements, and data safeguards.
Participate in joint pilot projects to support Ukraine’s ongoing transformation and share its valuable practices with the international community.
Improve AI performance monitoring and evaluation (Impact Monitoring)
As currently being carried out, the MDT could introduce systematic and structured mechanisms to monitor, evaluate, and improve AI-driven projects at both the organisational level and, over time, at the whole-of-government level.
Use AI tools to analyse user feedback from surveys and chatbot interactions in real time
Adopt post-deployment monitoring through live performance dashboards, scaling from individual projects to an integrated, whole-of-government approach to track the effectiveness of AI systems, user uptake, and cost savings.
Establish ex-ante and ex-post evaluation frameworks that are standardised across agencies, allowing for consistent assessment of AI systems' effectiveness, risks, and benefits at both the organisational and national levels.
Invest in capacity-building programmes to equip civil servants with the skills and tools needed to monitor, interpret, and act on AI performance data across government, as well as to develop general AI skills for public servants as outlined above.
Recovery and reconstruction priorities
Formalise Ukraine’s governance arrangements around AI in the public sector (Enabler)
Ukraine’s centre of government could establish a permanent coordination mechanism, refine roles and responsibilities of CDTOs, and further develop a strategic approach to digital government investments.
Establish a permanent coordination mechanism, coordinated by the MDT with participation of key stakeholders from the public sector and private sector representatives, at the centre of government dedicated to AI to fortify setting, implementing and monitoring of Ukraine’s strategic approach to AI over time, ensuring continuity of ongoing and future initiatives. This approach would also present an opportunity to achieve greater policy coherence, develop common policy tools such as standards and guidelines, and strengthen accountability through systematically monitoring and assessing the impact of AI.
Refine and strengthen the roles and responsibilities of the CDTOs to enhance implementation capacity and institutionalise accountability across the public sector (short-term).
Consider developing a strategic approach to digital government investments, including AI initiatives to ensure its resources are invested in line with the national strategies and long-term priorities.
Establish structured and innovation procurement for AI (Guardrail)
The Public Procurement Agency could establish dedicated AI processes and promote innovation procurement through the upcoming GovTech programme and procurement guidelines.
Launch small-scale AI pilot projects to refine procurement processes and evaluation criteria.
Co-develop procurement guidelines with tech companies, academia, and civil society.
Enhance the capacity of procurement officials through training on AI literacy, accountability, and transparency.
Promote innovation procurement through the upcoming GovTech programme by supporting trustworthy AI-driven solutions from startups and SMEs to address challenges in service delivery.
Enhance oversight mechanisms (Guardrail)
The Government of Ukraine could ensure long-term and sustained oversight of AI developments in government by continuing the efforts to establish an independent AI regulatory body, as planned with the future unified digital regulator, along with further developing the MDT monitoring function.
Continue working on the creation of an independent AI regulatory body via a participatory process involving experts and stakeholders in alignment with the EU AI Act.
At the executive level, develop the MDT AI monitoring function with tools and ensure consistent mapping of AI solutions, alignment with ethical standards, and risk mitigation strategies
Long-term development priorities
Enhance Ukraine’s digital public infrastructure including government data (Enabler)
Ukraine should continue to invest in key digital infrastructure to establish a set of shared, secure and interoperable digital systems.
Continue enhancing Trembita, focusing on improving data standards, interoperability across the public sector, and work towards achieving real-time access across registries and systems.
Consider taking a Government as a Platform (GaaP) approach more widely through investing in digital public infrastructure (DPI).
Anchor the WINWIN AI Centre of Excellence as a collaborative innovation hub (Engagement)
Monitor the Centre’s activities and report its progress and performance against defined objectives transparently to the public
Create a community of practice for the Ukrainian and international public sectors to share use cases and lessons learned.
Strengthen monitoring of AI experimentation (Impact Monitoring)
The MDT could support responsible AI experimentation by systematically managing experiments to ensure sustained scalability and public trust, both within the WINWIN AI Centre of Excellence and more broadly across the public administration, as exemplified by its ongoing bottom-up idea-collection efforts.
Define clear entry/exit criteria for AI pilots within the sandbox or a dedicated incubator.
Evaluate pilot projects using technical, ethical, and service quality indicators with end-user feedback.
Involve users and experts in the experimentation process.
Share lessons learned publicly and ensure interdisciplinary oversight from legal, technical, and civil society experts.
Ukraine has transitioned from an early endorsement of ethical AI frameworks to active participation in global AI safety coalitions, progressing to the signing of binding legal standards through the Council of Europe. This highlights Ukraine’s approach, from embracing global norms to fostering international safety collaboration and implementing enforceable legal commitments, in promoting the responsible use of AI in government and society. In parallel, Ukraine is preparing to align with the European Union’s Artificial Intelligence Act, which is expected to be adopted by 2027, further anchoring its regulatory framework within the European and international standards (OECD, 2025[23]).
OECD Recommendation on Artificial Intelligence: Ukraine has adhered to the OECD Recommendation on Artificial Intelligence [OECD/LEGAL/0449], including the OECD AI Principles, first adopted in 2019 and updated in 2024 (OECD, 2024[32]). These foundational OECD legal instruments advocate for trustworthy, human-centric AI that respects human rights, democratic values, environmental considerations, fairness, transparency, and accountability (OECD, 2024[32]).
UNESCO’s Recommendation on the Ethics of Artificial Intelligence: In November 2021, UNESCO’s Recommendation on the Ethics of Artificial Intelligence was adopted by all 194 Member States, establishing a global framework centred on human rights, fairness, transparency, environmental considerations, and inclusive governance (UNESCO, n.d.[82]).
The Bletchley Declaration on AI Safety: In November 2023, Ukraine continued to foster international collaboration by signing the Bletchley Declaration at the UK-led AI Safety Summit (UK Government, 2025[83]). This initiative, backed by 28 countries including Ukraine, aims to enhance safety around frontier AI through shared risk assessments, scientific co-operation, and transparent oversight (UK Government, 2025[83]).
Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet: In February 2025, Ukraine joined the Paris AI Action Summit, signing the “Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet” (Government of France, 2025[84]). Over 50 countries, including Ukraine, endorsed AI core values such as openness, ethical governance, accessibility, labour-market impact, and environmental considerations (Government of France, 2025[84]).
Council of Europe’s Framework Convention on AI, Human Rights, Democracy, and the Rule of Law: In May 2025, Ukraine deepened its commitment by supporting the development and signing the Council of Europe (CoE) Framework Convention on AI, Human Rights, Democracy, and the Rule of Law in Strasbourg (Council of Europe, 2024[25]). As the world’s first legally binding treaty focused on AI governance, it aims to ensure that activities within the lifecycle of artificial intelligence systems are fully consistent with human rights, democracy and the rule of law, while being conducive to technological progress and innovation (Council of Europe, 2024[25]). To support the implementation of the Convention, Ukraine and the CoE have launched targeted capacity-building initiatives, including a specialised training course for public servants on the intersection of AI and human rights (Council of Europe, 2025[85]).
References
[16] Australian Goverment (2021), C2021G00267 - Order to identify the Minister responsible for the Digital Transformation Agency, and to Specify Functions for the Digital Transformation Agency, https://www.legislation.gov.au/C2021G00267/latest/text.
[41] Committee on AI in Ukraine (2023), A road map of AI regulation in Ukraine, https://ai.org.ua/a-road-map-of-ai-regulation-in-ukraine/.
[85] Council of Europe (2025), Council of Europe HELP course on Artificial Intelligence and Human Rights launched for public servants, https://www.coe.int/en/web/kyiv/-/council-of-europe-help-course-on-artificial-intelligence-and-human-rights-launched-for-public-servants.
[25] Council of Europe (2024), The Framework Convention on Artificial Intelligence (2024), https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence.
[57] Department for Science, Innovation and Technology (2024), GOV.UK, https://www.gov.uk/government/publications/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024.
[35] Digital Security Lab Ukraine (2024), Running Up That Hill: Artificial Intelligence in Ukrainian Public Sector, https://dslua.org/publications/running-up-that-hill-artificial-intelligence-in-ukrainian-public-sector/.
[17] Digital Transformation Agency of Australia (2025), Digital Transformation Agency, https://www.dta.gov.au/advice/digital-and-ict-investments/digital-and-ict-investment-oversight-framework-iof.
[19] Du, M. (2023), “Machine vs. human, who makes a better judgment on innovation? Take GPT-4 for example”, Frontiers in Artificial Intelligence, Vol. 6, https://doi.org/10.3389/frai.2023.1206516.
[72] East Europe Foundation (2023), Diia.Education platform, https://eef.org.ua/en/platforma-diya-osvita/.
[9] EU4Digital (2025), Ukraine Launches WINWIN AI Center to Drive Innovation, https://eufordigital.eu/ukraine-launches-winwin-ai-center-to-drive-innovation/.
[33] EU4Digital (2024), Human Rights in the Era of Artificial Intelligence: Challenges and Legal Regulation, https://eufordigital.eu/wp-content/uploads/2024/03/EU4DigitalUA-Human-rights-in-the-area-of-AI.pdf.
[79] EU4Digital (2024), Ukraine approves resolution for launch of ‘Mriya’ educational app, https://eufordigital.eu/ukraine-approves-resolution-for-launch-of-mriya-educational-app/.
[77] EU4DigitalUA (2021), Diia mobile application - Evaluation Report - E-Governance Academy, https://eu4digitalua.eu/wp-content/uploads/2021/12/diia-evaluation-report.pdf.
[58] Euporean Union (2025), EU4Digital, https://eufordigital.eu/ukraine-launches-winwin-ai-center-to-drive-innovation/.
[31] Europe Union (2019), Ethics guidelines for trustworthy AI, https://op.europa.eu/en/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1.
[47] European Commission (2024), Competencies and governance practices for AI in the public sector, Joint Research Centre, Publications Office of the European Union, https://data.europa.eu/doi/10.2760/7895569.
[45] European Commission (2024), Ukraine 2024 Report - Commission Staff Working Document, https://enlargement.ec.europa.eu/document/download/1924a044-b30f-48a2-99c1-50edeac14da1_en.
[56] European Union (2024), EU4Digital, https://eufordigital.eu/ukraine-approves-resolution-for-launch-of-mriya-educational-app/.
[27] European Union (2024), Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1689.
[78] European Union (n.d.), SkillLab: Supporting the people and employers of Ukraine, European Year of Skills, https://year-of-skills.europa.eu/projects/skilllab-supporting-people-and-employers-ukraine_en?prefLang=sk\.
[84] Government of France (2025), Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet, https://www.elysee.fr/en/emmanuel-macron/2025/02/11/statement-on-inclusive-and-sustainable-artificial-intelligence-for-people-and-the-planet.
[52] Government of Singapore (2025), Public Sector AI Playbook.
[54] Government of Singapore (n.d.), Public Sector AI Playbook, https://www.developer.tech.gov.sg/products/collections/data-science-and-artificial-intelligence/playbooks/public-sector-ai-playbook.pdf.
[80] Government of Singapore (n.d.), Whole of Government Application Analytics (WOGAA), https://www.tech.gov.sg/products-and-services/for-government-agencies/data-and-ai/wogaa/.
[49] Government of Sweden (2025), Digg and IMY release national guidelines for the use of generative AI in public administration, https://www.digg.se/om-oss/nyheter/ai/nyheter/2025-01-21-digg-och-imy-slapper-nationella-riktlinjer-for-anvandningen-av-generativ-ai-inom-offentlig-forvaltning.
[48] Government of Sweden (n.d.), Guidelines for generative AI in public administration, https://www.digg.se/ai-for-offentlig-forvaltning/riktlinjer-for-generativ-ai.
[60] Government of the UK (2024), GOV.UK, https://www.gov.uk/service-manual/communities/artificial-intelligence-community.
[70] Government of Ukraine (2025), An AI assistant for receiving services will appear in Diia - Diia Web Portal, https://diia.gov.ua/news/u-dii-ziavytsia-shi-asystent-dlia-otrymannia-posluh.
[39] Government of Ukraine (2025), Diia - An AI assistant for receiving services will appear in Diia, https://diia.gov.ua/news/u-dii-ziavytsia-shi-asystent-dlia-otrymannia-posluh.
[71] Government of Ukraine (2025), Diia Education - Artificial Intelligence, https://osvita.diia.gov.ua/en/catalog/topic/artificial-intelligence.
[69] Government of Ukraine (2025), Mriia: Ukraine’s Innovative Educational Ecosystem, https://digitalstate.gov.ua/news/govtech/mriia-innovatsiyna-osvitnia-ekosystema-ukrayiny.
[62] Government of Ukraine (2025), The Ministry of Digital Transformation launches the WINWIN AI Center of Excellence — a center of excellence in AI development and integration, Press Office of the Ministry of Digital Transformation of Ukraine, https://thedigital.gov.ua/news/mintsifra-zapuskae-winwin-ai-center-of-excellence-tsentr-peredovogo-dosvidu-z-rozrobki-ta-integratsii-shi.
[36] Government of Ukraine (2025), The State Service for Special Communications will train in the use of artificial intelligence in the field of cybersecurity, https://cip.gov.ua/ua/news/u-derzhspeczv-yazku-navchatimut-vikoristannyu-shtuchnogo-intelektu-v-sferi-kiberbezpeki.
[34] Government of Ukraine (2025), Tips for the Responsible Use of Artificial Intelligence by Public Servants - Ministry of Digital Transformation of Ukraine, https://cms.thedigital.gov.ua/storage/uploads/files/page/community/docs/Porady_z_vykorystannya_AI.pdf.
[73] Government of Ukraine (2025), Ukraine Joins Global AI Race with Its Own Language Model, https://digitalstate.gov.ua/news/govtech/start-rozrobky-ukrayinskoyi-llm-partnerstvo-mintsyfry-ta-kyyivstaru.
[4] Government of Ukraine (2025), Ukraine’s Diia Hailed as «Stunning» Digital Service by Lisbon Council, https://digitalstate.gov.ua/news/govtech/ukraines-diia-hailed-as-stunning-digital-service-by-lisbon-council.
[30] Government of Ukraine (2025), Ukraine’s Global Innovation Strategy 2030, https://winwin.gov.ua/en.
[66] Government of Ukraine (2025), We are developing the Ukrainian Large Language Model (LLM) — improving government services, supporting business, and stepping towards leadership in AI, https://thedigital.gov.ua/news/rozroblyaemo-ukrainsku-veliku-movnu-model-llm-pokrashchuemo-derzhavni-servisi-pidtrimuemo-biznes-i-krokuemo-do-liderstva-v-shi.
[29] Government of Ukraine (2024), Concept of the State Targeted Scientific and Technical Program for the Use of Artificial Intelligence Technologies in Priority Economic Sectors until 2026, https://zakon.rada.gov.ua/laws/show/320-2024-%D1%80#Text.
[26] Government of Ukraine (2024), Glossary of Terms on AI - Ministry of Digital Transformation, https://thedigital.gov.ua/storage/uploads/files/page/community/docs/%D0%A1%D0%BB%D0%BE%D0%B2%D0%BD%D0%B8%D0%BA%20%D1%82%D0%B5%D1%80%D0%BC%D1%96%D0%BD%D1%96%D0%B2%20%D1%83%20%D1%81%D1%84%D0%B5%D1%80%D1%96%20%D0%A8%D0%86%20(1).pdf.
[6] Government of Ukraine (2024), Strategy for the Digital Development of Innovation Activity of Ukraine for the period up to 2030, https://zakon.rada.gov.ua/laws/show/1351-2024-%D1%80/print.
[64] Government of Ukraine (2024), The Ministry of Foreign Affairs of Ukraine has appointed a digital person to provide information on consular issues, https://mfa.gov.ua/news/mzs-ukrayini-priznachilo-cifrovu-osobu-dlya-informuvannya-shchodo-konsulskih-pitan.
[22] Government of Ukraine (2024), Ukraine’s AI regulation approach - Ministry of Digital Transformation, https://webportal.nrada.gov.ua/wp-content/uploads/2024/12/02_Dubno_AI-regulation-11_12.pdf.
[7] Government of Ukraine (2024), White Paper on Artificial Intelligence Regulation in Ukraine: Vision of the Ministry of Digital Transformation of Ukraine - Version for Consultation, https://thedigital.gov.ua/storage/uploads/files/page/community/docs/%D0%91%D1%96%D0%BB%D0%B0_%D0%BA%D0%BD%D0%B8%D0%B3%D0%B0_%D0%B7_%D1%80%D0%B5%D0%B3%D1%83%D0%BB%D1%8E%D0%B2%D0%B0%D0%BD%D0%BD%D1%8F_%D0%A8%D0%86_%D0%B2_%D0%A3%D0%BA%D1%80%D0%B0%D1%97%D0%BD%.
[12] Government of Ukraine (2023), CDTO Campus, https://www.cdto-campus.com/en.
[24] Government of Ukraine (2023), Regulation of artificial intelligence in Ukraine: Ministry of Digital Transformation of Ukraine presented a roadmap, https://www.kmu.gov.ua/news/rehuliuvannia-shtuchnoho-intelektu-v-ukraini-mintsyfry-prezentuvalo-dorozhniu-kartu.
[5] Government of Ukraine (2020), Concept for the development of artificial intelligence in Ukraine, https://zakon.rada.gov.ua/laws/show/1556-2020-%D1%80/print.
[28] IFES (2024), Adapting EU Artificial Intelligence Regulations for Electoral Processes: A Path for Ukraine, https://www.ifesukraine.org/wp-content/uploads/2024/09/ifes-artificial-intelligence-eng-5.pdf.
[14] Korean Government (2024), National Artificial Intelligence Committee, https://aikorea.go.kr/web/main.do.
[68] Limantė, A. and Y. Moskvytyn (2024), Integrating Artificial Intelligence in Ukraine’s Courts: State of Play and Future Prospect,, VerfBlog, 2024/12/12, https://doi.org/10.59704/c472a6ec2910dae3.
[21] Ludwig, J. (2022), Algorithmic behavioral science: Machine learning as a tool for scientific discovery.
[20] Mac Aonghusa, P. and S. Michie (2020), “Artificial Intelligence and Behavioral Science Through the Looking Glass: Challenges for Real-World Application”, Annals of Behavioral Medicine, Vol. 54/12, pp. 942-947, https://doi.org/10.1093/abm/kaaa095.
[13] Ministry of Digital Transformation (2025), Diia.Education, https://osvita.diia.gov.ua/en.
[8] Ministry of Digital Transformation (2025), Ministry Press Office, https://thedigital.gov.ua/news/mintsifra-zapuskae-winwin-ai-center-of-excellence-tsentr-peredovogo-dosvidu-z-rozrobki-ta-integratsii-shi.
[50] Monteiro, B., A. Hlacs and P. Boéchat (2024), Public procurement for public sector innovation: Facilitating innovators’ access to innovation procurement, OECD Working Papers on Public Governance, No. 80, OECD Publishing, Paris,, https://doi.org/10.1787/9.
[10] National Agency of Ukraine on Civil Service (2025), Guidelines on Responsible Use of Artificial Intelligence by Public Servants, https://nads.gov.ua/ (accessed on 30 March 2025).
[11] National Agency of Ukraine on Civil Service (2024), National Agency of Ukraine on Civil Service, https://pdp.nacs.gov.ua/courses/zastosuvannia-shtuchnoho-intelektu-funktsional-ta-osoblyvosti.
[15] OECD (2025), Effectively managing investments in digital government: The Digital Government Investment Framework, https://doi.org/10.1787/5c324e91-en.
[1] OECD (2025), Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Function, OECD Publishing, Paris, https://doi.org/10.1787/795de142-en.
[44] OECD (2025), Survey on Digital Government 3.0.
[37] OECD (2025), Towards a common reporting framework for AI incidents, OECD Artificial Intelligence Papers, No. 34, OECD Publishing, Paris, https://doi.org/10.1787/f326d4ac-en.
[23] OECD (2025), Ukraine’s use of AI for service design and delivery, and the public sector - Interviews, Public Governance Review of Ukraine.
[18] OECD (2024), “Digital public infrastructure for digital governments”, OECD Public Governance Policy Papers, No. 68, OECD Publishing, Paris, https://doi.org/10.1787/ff525dc8-en.
[51] OECD (2024), Enabling Digital Innovation in Government: The OECD GovTech Policy Framework, OECD Digital Government Studies, OECD Publishing, Paris,, https://doi.org/10.1787/a51eb9b2-en.
[2] OECD (2024), Governing with Artificial Intelligence: Are governments ready?, OECD Artificial Intelligence Papers, No. 20, OECD Publishing, Paris, https://doi.org/10.1787/26324bc2-en.
[42] OECD (2024), OECD AI principles, https://www.oecd.org/en/topics/ai-principles.html.
[3] OECD (2024), Public Administration in Ukraine: Assessment against the Principles of Public Administration, SIGMA Monitoring Reports, OECD Publishing, Paris, https://doi.org/10.1787/078d08d4-en.
[32] OECD (2024), Recommendation of the Council on Artificial Intelligence, https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449.
[61] OECD (2024), Recommendation of the Council on Human-Centred Public Administrative Services, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0503.
[43] OECD (2023), National Council of Ukraine on the Development of Science and Technology - STIP Compass, https://stip.oecd.org/stip/interactive-dashboards/policy-initiatives/2023%2Fdata%2FpolicyInitiatives%2F27074.
[38] OECD (n.d.), OECD AI Incidents Monitor (AIM) - OECD AI Policy Observatory, https://oecd.ai/en/incidents.
[40] OECD/UNESCO (2024), G7 Toolkit for Artificial Intelligence in the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/421c1244-en.
[55] papers, O. (ed.) (forthcoming), The use of AI across the innovation process in government.
[75] SkillLab (2025), Unlocking the potential of Ukraine’s workforce, https://skilllab.io/en-us/news/unlocking-workforce-potential-in-ukraine.
[74] The Odessa Journal (2025), 5 AI Products from the Ministry of Digital Transformation, https://odessa-journal.com/5-ai-products-from-the-ministry-of-digital-transformation.
[83] UK Government (2025), The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023 - Policy paper, GOV.UK, https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023.
[46] UK Government (2024), The Model for Responsible Innovation, https://www.gov.uk/government/publications/the-model-for-responsible-innovation.
[53] UK Government (2020), Guidelines for AI procurement - GOV.UK, https://www.gov.uk/government/publications/guidelines-for-ai-procurement/guidelines-for-ai-procurement.
[81] UK Information Commissioner’s Office (2024), Regulatory Sandbox Insights - Report 2024, https://ico.org.uk/media/for-organisations/documents/4030434/regulatory-sandbox-insights-report.pdf.
[76] UNDP (2025), Accessibility of Ukrainian government websites increases for fourth year in a row, https://www.undp.org/ukraine/press-releases/accessibility-ukrainian-government-websites-increases-fourth-year-row.
[67] UNDP (2025), Ukrainians’ satisfaction with government e-services on the rise, KIIS survey finds, https://www.undp.org/ukraine/press-releases/ukrainians-satisfaction-government-e-services-rise-kiis-survey-finds#.
[65] UNDP (2024), Analytical report “Opinions and views of Ukrainians on state electronic services in 2024”, https://www.undp.org/ukraine/publications/analytical-report-opinions-and-views-ukrainians-state-electronic-services-2024.
[82] UNESCO (n.d.), Ethics of Artificial Intelligence - The Recommendation, https://www.unesco.org/en/artificial-intelligence/recommendation-ethics.
[59] United Nations Industrial Development Organization (2024), United Nations Industrial Development Organization, https://www.unido.org/news/investing-ukraine-unido-government-ukraine-and-european-union-collaborate-leverage-ai-ukraines-economic-recovery-and-europes-global-competitiveness.
[63] UNN (2024), AI-based chatbot to appear in Diia - Fedorov, https://unn.ua/en/news/ai-based-chatbot-to-appear-in-diia-fedorov.
← 1. A comprehensive review of cybersecurity practices was not carried out. The OECD Directorate for Science, Technology and Innovation (STI) addresses digital security through a dedicated workstream. See https://www.oecd.org/en/topics/policy-issues/digital-security for more information.