The Supreme Audit Office of the Slovak Republic (Najvyšší kontrolný úrad SR) is a large institution, and like most institutions its size, its people aren’t all starting from the same place with AI. Some staff, analysts among them, already had experience with AI, and others had none. Instead of one course for everyone, we put together several parallel approaches, each aimed at the group that actually needed it.
AI literacy is about understanding what AI actually is, what it can do, and the regulatory and ethical questions around using it. For most of the staff, that was the starting point. For officials handling documents and administrative work day to day, we adapted a basic prompting course to what they actually do. And for the analysts who were already using AI in their work, we went further, with more technical consultations built around their own tasks.
“Using AI effectively is not only about knowing what these tools can do, but also understanding how they work and where their limitations lie. As AI becomes more common in our work, it is important to be aware of the potential risks that can arise from using these tools without sufficient knowledge. Understanding how AI learns, processes information, and generates its outputs can help us avoid overreliance, put generated results into context, and critically assess their quality and reliability. I hope the AI Literacy course helped participants to look ‘under the hood’ and gain a clearer picture of what happens behind the outputs they see.”
Alongside the literacy track, we ran a separate set of consultations on AI regulation and ethics, aimed at their lawyers, among others. These sessions worked through practical questions about how the regulations apply to the everyday work of a public institution. One part covered the AI Act, the EU’s first regulation of artificial intelligence, adopted in 2024, working through which systems and organizations it applies to, how the EU grades AI systems and what it means in daily practice. Another part covered GDPR and the questions it still raises for AI, including what counts as personal data, how to determine a lawful basis for processing it, and whether GDPR’s research regime can be used for training AI.
“It is absolutely necessary for public authorities to understand the benefits and risks of AI deployment in the public sector. Moreover, basic understanding of AI and related EU laws is essential when it comes to deployment of AI tools in public institutions. I truly believe that our courses will help employees of the Supreme Audit Office with successful implementation of lawful and trustworthy AI solutions.”
Beyond AI literacy and regulation, the collaboration also covered machine learning in practice, working with large documents using LLMs, AI in practice, personal data protection, and training and using custom language models without sharing data with external services, along with web scraping.
It’s this mix, AI literacy as a shared starting point for everyone, regulation for the people who need to navigate it, and deeper understanding for the ones already using AI, that we can put together for other public institutions and companies, too.
Jana Juriová
Head of Strategy and Analysis Department
