Slovakia’s MIRRI bill awaits Parliament before planned 1 January national AI framework
The proposal would place a National Centre for Artificial Intelligence and Data within MIRRI, combining statutory supervision with advice and business support.
Slovakia’s Ministry of Investments, Regional Development and Informatization (MIRRI) has prepared a government bill to create the national framework for applying the EU Artificial Intelligence Act and European data regulation. MIRRI said on 13 November 2026 that the bill had been approved by the relevant committees of the National Council and was awaiting second and third readings in Parliament. Its expected entry into force is 1 January 2027.
The proposed structure matters to compliance teams because Slovakia is signalling a supervisory model built around advice and prevention rather than immediate punishment. The bill would establish a National Centre for Artificial Intelligence and Data within MIRRI. The centre is intended both to exercise statutory supervisory powers and to provide information and support on applying the new rules. Until Parliament completes the bill’s remaining stages, however, the framework remains proposed rather than in force.
A support-led supervisory model
Radoslav Štefánek, the government’s plenipotentiary for artificial intelligence, said the national law had been designed as simply as possible. “Our goal is not to fine entrepreneurs, but to help them adapt to new rules,” he said, in a translation from Slovak. MIRRI said it would prioritise “advice, prevention and a proportionate approach to supervision”.
That approach is aimed particularly at small and medium-sized businesses. The stated objective is to help them understand their obligations under the European rules while allowing them to develop AI products for international markets. The planned centre therefore appears designed to serve as both a supervisory point of contact and an implementation resource for companies that need help interpreting the new framework.
The source does not set out the bill’s proposed allocation of specific enforcement powers, penalties or institutional responsibilities beyond the centre’s planned role. Compliance teams should therefore treat the public announcement as a policy and legislative update, not as a substitute for the enacted text.
The operational message for firms
Štefánek’s guidance identifies two immediate control concerns for companies using AI.
First, firms should verify AI outputs rather than treat them as authoritative. He said natural-language interaction makes users inclined to trust AI because “we talk to it like a colleague”, even though its outputs must be checked. The quotation is translated from Slovak.
Second, companies should protect sensitive business information and use tools whose data-processing and security conditions they understand. That points to practical review of vendor terms, data flows and safeguards before staff use external or internally deployed AI tools with company information.
MIRRI also links compliance to organisational readiness. Digital transformation, Štefánek said, involves more than buying software licences. It requires changes to work processes, employee skills and an organisation’s ability to use new technologies effectively. For businesses, the proposed national framework is therefore likely to sit alongside internal governance work covering permitted uses, staff capability and oversight of AI-enabled processes.
Atlas’s view is that the planned centre could make the quality of Slovak guidance as important as the formal supervisory powers. A support-led model may reduce uncertainty for smaller firms, but only if advice is sufficiently specific to help them map real deployments to their obligations.
A broader innovation policy
MIRRI’s announcement presents the national framework as part of a wider effort to develop Slovak AI capability. Štefánek said Slovakia should focus on practical applications that solve business problems and improve productivity, rather than compete with global technology companies on the largest language models.
The ministry also highlighted the value of Slovak-language data for adapting global models to the country’s linguistic, legal and cultural environment. It pointed to European cooperation, including EuroHPC initiatives, as a route to computing capacity and research infrastructure for researchers and innovative businesses.
For public administration, Štefánek sees potential for open language models adapted to state services and legal needs to process documents, speed up administrative work and reduce the burden on businesses. That use case would itself require the process, skills and data controls described by the plenipotentiary.
The bill’s next legislative stage is its second and third reading in the plenary of the National Council. MIRRI expects the law to enter into force on 1 January 2027.
Permanent link to this Atlas analysis
Related Atlas analysis
Continue with coverage connected by market, topic and operating context.
- BrazilBrazil’s ANPD chairs approval of Ibero-American AI data standardsThe guidance says public availability is not a licence to train models, and generic AI notices are inadequate.
- United StatesNevada proposes new pari-mutuel controls effective 1 October 2027The draft would require at least 26 report types, 60-day location recordings and no manual tickets during system failure.
- GreeceGreece enacts Law 5354 with CRD VI, crypto and AML changesThe 9 November law transfers operation of the beneficial ownership register to Greece’s tax authority and adds new controls for third-country bank branches.
Continue with Atlas
Move from this development into the relevant research, comparison and workflow.