Greece is moving to create a centralized registry of artificial intelligence systems used across the public sector, in a bid to bring greater visibility and oversight to the rapid adoption of AI by government agencies.
The measure is included in draft legislation implementing the European Union's AI Act, which was released for public consultation this week. At the center of the proposal is a requirement that every public-sector body—from ministries and state agencies to independent authorities—register any AI system before it enters operational use.
The registry would be maintained by the Special Secretariat for Artificial Intelligence and Data Governance within the Ministry of Digital Governance, creating for the first time a single inventory of AI applications deployed across the Greek state.
ΔΙΑΒΑΣΤΕ ΕΠΙΣΗΣ
Government officials view the initiative as a key governance tool at a time when public administrations across Europe are increasingly experimenting with AI-driven services. By collecting information on systems before they go live, authorities hope to gain a clearer picture of the government's digital capabilities and identify where artificial intelligence is being developed, procured or deployed.
Supporters of the measure argue that a centralized registry could help reduce duplication of projects across agencies, improve coordination of AI investments and facilitate the sharing of best practices throughout the public sector. They also see it as an early accountability mechanism for technologies that increasingly influence decisions affecting citizens and businesses.
Transparency Tool or Bureaucratic Burden?
Yet the proposal is already prompting debate over whether the registry will become a meaningful transparency tool or merely another layer of bureaucracy.
One of the main concerns centers on public access. While the draft legislation requires agencies to submit information about AI systems, it remains unclear which parts of that information will be publicly available and which will remain internal. The bill also does not fully explain how citizens will be informed when a public service or administrative decision relies on artificial intelligence.
Digital-governance specialists have raised additional questions about the impact on innovation. Because registration must take place before a system enters operation, critics warn that the process could slow pilot programs and experimental projects that depend on rapid development cycles. Such concerns are particularly relevant as governments across Europe seek to accelerate AI adoption while simultaneously introducing new compliance obligations under the EU's regulatory framework.
Another challenge involves the breadth of technologies that could fall under the registry's scope. Without a clearly defined risk-classification framework, experts caution that the database could end up grouping together everything from simple automation tools and chatbots to sophisticated decision-support systems. Such an approach, they argue, may make it difficult for regulators to distinguish between low-risk applications and systems that warrant closer scrutiny.
Perhaps the most significant concern relates to verification. Under the current draft, the information entered into the registry would be submitted by the same public bodies deploying the AI systems. The legislation does not specify an independent auditing mechanism to verify the accuracy, completeness or consistency of those disclosures.
The debate reflects a broader challenge facing governments across Europe as they begin translating the EU AI Act into national governance structures. Policymakers are seeking to balance innovation with accountability, while ensuring that transparency measures do not become compliance exercises with limited practical value.
