Regulation of artificial intelligence: Ukraine’s current approach

Oleksandr Kozhukhar

AVELLUM, Kyiv

okozhukhar@avellum.com

Mykhailo Petrenko

AVELLUM, Kyiv

mpetrenko@avellum.com

General overview

As Ukraine progresses towards full integration into the European Union’s digital single market, one of its strategic objectives is the harmonisation of the national legal framework with the EU’s AI Act. However, given the rapid pace of technological evolution, the specific constraints of the current environment and some uncertainty around the AI Act itself (in view of the EU Omnibus simplification package),[1] the Ministry of Digital Transformation has adopted a pragmatic, ‘bottom-up’ approach. This is detailed in the White Paper on AI Regulation in Ukraine (June 2024), prioritising gradual alignment over immediate, rigid legislation.

This strategy is designed to balance two competing interests: the protection of human rights and the fostering of innovation. By avoiding premature strict regulation, the state aims to provide businesses with the necessary time and tools to prepare for future compliance requirements, thereby supporting the competitiveness of Ukrainian enterprises in global markets.

The two-stage regulatory concept

The current approach divides the pathway towards mandatory regulation into two distinct phases: a preparatory phase using soft law instruments; and a legislative implementation phase.

Stage one: preparatory phase

The initial stage is planned to span approximately two to three years, focusing on building the capacity of both the state and the private sector to manage AI risks effectively. Key instruments introduced during this phase include:

Regulatory sandbox

Regulatory sandbox (testing) mechanism constitutes an experimental project coordinated by the Ministry of Digital Transformation. It provides a controlled environment where legal entities can conduct research into high-tech solutions – specifically those using AI and blockchain – under state supervision. The Innovation Development Fund acts as the organiser, providing information support and consulting to participants to assess product safety, legal compliance and intellectual property rights before market release.

Voluntary Code of Conduct

Marking a move towards ethical self-regulation, the Voluntary Code of Conduct on the Ethical and Responsible Use of AI was endorsed in December 2024. Signatories, comprising a number of leading Ukrainian technology companies, have voluntarily committed to adhering to unified ethical principles, including transparency, fairness, safety and data protection. Participating entities undertake to implement risk management systems and report on their adherence to the Code’s provisions.

Sectoral recommendations

To provide specialised guidance, the Ministry has issued a series of recommendation documents serving as soft law for specific industries. These instruments cover diverse spheres, ranging from guidelines for the media sector and educational institutions to specialised protocols for legal professionals.

Assessment methodology

The White Paper provides for the introduction of a methodology for assessing the impact of AI systems on human rights. Grounded in the Council of Europe’s HUDERIA framework, this tool enables the categorisation of products by risk level (low, medium or high). By applying a risk-based approach, developers can independently evaluate the potential impact of their systems, facilitating preparation for future regulatory requirements.

Voluntary labelling

The White Paper introduces a transparency mechanism involves the voluntary disclosure by developers of information regarding system architecture, training datasets and decision-making algorithms. This mechanism is intended to provide consumers with clear information similar to product labelling.

Platform for legal aid

The White Paper also anticipates the creation of a Platform for legal aid. This resource is designed to provide consultation support to developers regarding the compliance of their products with future regulatory requirements. This instrument is primarily targeted at entities that do not participate in the regulatory sandbox but require expertise to ensure compliance.

Stage two: legislative alignment

The second stage involves the adoption of national legislation similar to the EU AI Act. The White Paper sets out the key structural elements of this future regulation:

• in alignment with the EU regulatory framework, the scope of obligations will be determined by the risk level that an AI system poses to human rights and safety;
• the law is expected to include a deferred entry-into-force mechanism for specific provisions, providing a transitional period for market adaptation similar to the timeline established in the EU AI Act; and
• the implementation of the law entails the establishment of a competent state authority or authorities responsible for market surveillance and the enforcement of sanctions.

Scope and exemptions

Defence sector exemption

An important feature of the Ukrainian approach, which aligns with the EU AI Act’s exemptions, is the exclusion of the defence sector from this regulatory framework, particularly with due regard to the ongoing martial law. Restrictions on the use of AI for defence purposes are viewed to disadvantage Ukraine relative to the aggressor state. Regulation in this domain is limited to international humanitarian law considerations.

Interplay with existing laws

The White Paper clarifies that the absence of a specific AI law during stage one does not exempt entities from existing legal obligations. Issues such as intellectual property rights regarding AI-generated content and the protection of personal data remain governed by current sectoral laws.

Cooperation with technology platforms

The Ministry of Digital Transformation proposes a partnership model to address rights violations on digital platforms. The framework empowers reputable NGOs to review complaints regarding AI-related human rights violations and escalate them to platforms for consideration.

Final remarks

Ukraine’s approach to AI regulation prioritises alignment with EU standards while addressing local security and market constraints. The initial focus on soft law and voluntary compliance aims to build necessary capacity prior to the enforcement of mandatory statutory requirements.