Greece Moves to Implement National AI Act Framework

A Greek parliamentary committee has opened debate on a national bill implementing the EU's Artificial Intelligence Act, known as the AI Act. The draft, prepared by the Ministry of Digital Governance, has already gone through public consultation and aims to establish a national oversight system for AI covering public administration, business and citizens.

What changes for citizens

Under the draft, citizens gain greater clarity about when they are dealing with an AI system versus a human. AI systems used in recruitment, education or other decisions with significant impact on citizens' lives will face additional oversight. The bill also provides for a single, unified channel for complaints about AI systems, replacing the scattered procedures currently spread across individual agencies.

The approach mirrors measures being rolled out in parallel across other EU member states, including Poland, where from August 2 companies offering chatbots will have to inform users they are talking to a machine. Greece is going a step further, building an entire institutional architecture for AI oversight from scratch rather than just a single disclosure requirement.

A sandbox for startups and business

For businesses, the most important element of the draft is the promised stability and predictability of the regulatory environment under uniform EU rules. The regulatory sandbox included in the bill is meant to let startups and small and medium-sized enterprises test innovative AI solutions with state institutional support before they reach the market, reducing legal uncertainty and encouraging investment in the sector.

The new authority, the Coordination and Expertise Centre, is meant to serve as a central point coordinating AI oversight across public administration, health, education and employment, rather than leaving that responsibility split among scattered sectoral institutions, as is currently the case in many EU countries.

The discussion about artificial intelligence isn't just about technology. Above all, it's about trust - Dimitris Papastergiou, Greece's Minister of Digital Governance

AI registry in public administration

One of the more practical elements of the draft is a unified registry that will record every AI system used by public institutions. It responds to a problem that agencies themselves have flagged in many EU countries, including Poland: a lack of full visibility into where and to what extent public administration is already using AI, which makes enforcing obligations under the EU regulation harder.

The draft moved to the parliamentary committee after public consultation wrapped up in early July. The government had said it would send the bill to parliament as early as June, but a final vote has not yet been scheduled.

What it means for Poland

The Greek model illustrates one path member states can take when implementing the AI Act: building a dedicated coordinating authority instead of splitting competencies among existing institutions. That question remains open in Poland too, where work continues on national implementing legislation for the EU regulation and where agencies themselves admit they don't yet have a full picture of their own AI use.

For Polish tech companies operating in the Greek market or planning to expand in the region, this means tracking not just EU rules but also national implementing measures, which can differ significantly between member states in the scope of obligations and the structure of oversight.

Sources: ProtoThema English (en.protothema.gr), Mondaq (mondaq.com)