Croatia Has Not Yet Selected a National AI Regulator

European Union member states were required to inform the European Commission by August 2 of this year about which national body, or bodies, would be responsible for the implementation and oversight of the EU Regulation on Artificial Intelligence, or the AI Act. Most countries have not yet done so, including Croatia, which means that currently no one is overseeing the enforcement of the rules or sanctioning irregularities, even though the AI Act introduces strict rules for the development and use of artificial intelligence, especially for high-risk systems.

Although the EU has established the obligation to appoint competent national authorities, member states are lagging to the extent that Brussels has not yet published an official list of national supervisory bodies. It is only known which countries have established a so-called single contact point for communication regarding the application of the AI Regulation, and there are only seven: Cyprus, Ireland, Latvia, Lithuania, Luxembourg, Slovenia, and Spain. In Slovenia, for example, this role has been taken on by the Agency for Communication Networks and Services.

The Croatian Ministry of Justice, Administration and Digital Transformation has so far submitted only a preliminary list of ‘competent authorities’ to the European Commission, including the Agency for Personal Data Protection (AZOP), the Ombudsman, the Ombudsman for Children, the Ombudsman for Persons with Disabilities, the State Electoral Commission (DIP), and the Agency for Electronic Media (AEM).

In response to our inquiry about whether Croatia has selected a national regulator for artificial intelligence, the Ministry only states that the expert working group ‘is working on the preparation and establishment of a national institutional framework for the implementation of the AI Regulation.’ The working group includes representatives from state bodies, the academic community, industry, and civil society, and a draft law is currently being prepared to regulate the structure responsible for the field of artificial intelligence, and thus the national competent authority. The Ministry states that it regularly monitors the implementation of the Regulation and coordinates activities with the European Commission.

Is the digital omnibus the solution?

Given that most member states are lagging in the obligation to appoint a national AI regulator, the European Commission recently proposed the introduction of a so-called digital omnibus, a set of rules that would enable faster and simpler implementation of digital regulations. As we recently wrote in Lider, the aim of the omnibus is to facilitate access to data for the development of artificial intelligence for companies, especially smaller ones, while partially changing the provisions of the GDPR.

The digital omnibus gives primacy to the Commission in the implementation of the AI Regulation, explains lawyer Marijana Šarolić Robić, as member states ‘are significantly behind in implementing their obligations to appoint and determine the competencies of national AI regulators.’

However, in her opinion, this is not necessarily a bad thing.

– It is better to have one regulator at the EU level than 27 regulatory bodies with potentially different approaches to the application of the AI Regulation – says the lawyer.

However, the omnibus has not yet been adopted, and its voting is expected only in 2026, so until then, only the AI Regulation is in force. And while the Commission waits for national plans and member states finalize their laws, the oversight system for artificial intelligence in the EU currently has no clearly defined bodies or procedures.