Slovakia is preparing a vision for artificial intelligence by December 18, to be followed next year by a strategy. The documents are meant to set the direction and concrete steps to bring less bureaucracy, a safer online space for children, and a stronger economy. The vision will be published online for public comment so the public can also have a say.
From Vision to Strategy: What, When, and Why
The vision states where the country wants to head in AI; the strategy describes how to get there. That is why the vision will come first, followed by the strategy, which must undergo a lengthy interministerial review process. On December 18 there will be a brief press conference, and then the document will be posted online, where the public can submit comments. It is an open framework, not a finished regulation.
All central government bodies and other institutions, including the nuclear regulatory authority, collaborated on the preparation, since energy is key when planning data centers and so-called AI factories. According to the government envoy for AI, Radoslav Štefánek, it is important to understand what such infrastructure means for the country and its know-how. Slovak experts are in demand worldwide and are involved in building large AI centers abroad as well. The vision aims to unify the state’s perspective and set common rules of the game.
Three Priorities: Infrastructure, Data, and Education
The foundation is modern infrastructure and high-quality, verified data, ideally managed on domestic servers. The internet is increasingly filled with AI-generated content, which raises the risk of models “learning” from less reliable sources. An example of the advantage of quality data is when Google digitized the Austrian National Library—the models then draw on accurate, curated texts rather than the random web. Slovakia wants to apply a similar approach to decision-making in public services.
The second priority is a more automated public administration that will bring less bureaucracy, but decision-making will remain in a “human in the loop” mode—with a human at the end of the process. The third area is education: of civil servants, so they know how to use AI correctly and safely, and of companies, which today tend to be cautious in adopting tools. The goal is to expand the use of AI so that it is accurate, safe, and useful. That also includes a simple rule: sensitive documents do not belong in publicly available chatbots.
Obstacles and How to Overcome Them
Štefánek sees departmentalism and fragmentation as the biggest problem in the state—AI needs a unified, coordinated approach. Next is low data quality, because precise automation cannot be built on poor inputs. Fear of change is natural, but AI is above all a tool: it will replace mainly routine steps, while key decisions remain with people. The rules are also meant to prevent security failures, such as the use of a public model to analyze a sensitive document at one office.
Another condition is technological sovereignty: Europe has decided to strengthen its own solutions so it is not dependent on partners who could “switch us off” at a critical moment. Offers from large foreign vendors are welcome, but the state is also looking for quality European alternatives. A combination of centralized management, reliable data, and responsible use of tools is meant to turn the vision into practice. The result should be faster state services, a safer internet for children, and a more competitive economy.
