Opinion & Analysis

Artificial intelligence conformity assessment in Ukraine

Ukraine’s commitment to joining the EU shapes its approach to AI regulation. The EU AI Act makes conformity assessment the operational core of AI regulation and, as part of the EU acquis, a conditionality element attached to Ukraine’s accession and the Ukraine Facility. With the Digital Omnibus on AI deferring high-risk obligations to December 2027 and August 2028, Ukraine has an additional 12- to 16-month window to prepare its AI framework to match the EU.

This report examines what EU-compatible AI conformity assessment requires and where Ukraine stands. In the EU regime, most high-risk obligations rely on provider-side internal control rather than third-party certification, which makes harmonised standards from CEN-CENELEC JTC 21 central. Ukraine is already tracking these standards for direct implementation. Its staged strategy of voluntary preparation before binding legislation is sound, and its technical-regulation, accreditation and standardisation infrastructure provides a usable backbone for adaptation. The main gaps however are institutional and in expertise: Ukraine has no independent data-protection authority and no designated AI supervisor, the Single Digital Regulator may be delayed, and AI audit expertise is scarce. War and an unclear civilian/dual-use boundary also create unique challenges for conformity.

The report recommends using 2026–27 to build enforceability before legislating. The steps it proposes include completing data-protection reform, deepening standards alignment, developing AI audit skills, preparing a fallback supervisory option, mapping sectoral readiness, and organising alignment as an implementation partnership with EU bodies, especially in light of upcoming EU legislative initiatives.

About the authors:

César Dro is a EU policy professional with 7+ years of experience shaping research & innovation strategies at the European Commission. My work bridges foresight, policy analysis, and technology governance, with a focus on emerging and disruptive technologies.

Eduardo Brito is a Security Engineer and Researcher at Cybernetica AS, Industrial PhD candidate in Computer Science at the University of Tartu, and Visiting Researcher at CEPS and TU Berlin.

 

Access the original publication here