Slovakia - AI Ethics Commission (2020)

Overview

The Standing Commission for Ethics and Regulation of Artificial Intelligence (CERAI) was established as an independent expert and advisory body within the Ministry of Investments, Regional Development and Informatization of the Slovak Republic to address ethical, legal and societal issues linked to AI. Formally created on 2 November 2020, with membership appointments announced in early November 2020, CERAI commenced its practical activities in January 2021. The commission’s public profile and core documentation are hosted by the ministry; see the official commission page at MIRRI – CERAI. CERAI’s role is advisory: it provides expert opinions on draft legislation, issues ethical guidance for public sector AI deployment and supports cross-sectoral dialogue to foster human-centric, rights-respecting AI systems in Slovakia.

Definitions

Key definitions used by CERAI align with European practice and national law. "Artificial Intelligence" is understood broadly to include systems that perform tasks using data-driven or rule-based models, including machine learning models, deep learning neural networks, and hybrid approaches. "High‑risk AI" refers to deployments that materially affect safety, fundamental rights, or essential services. "Accountability documentation" includes model cards, datasheets for datasets, impact assessments and records of training/validation pipelines. "Transparent" means providing meaningful, context-appropriate explanations of system behaviour to affected stakeholders; "privacy" is interpreted in line with national data protection norms and the GDPR. The commission encourages consistent terminology across public procurement, regulatory guidance and sectoral policies so stakeholders can apply interoperable safeguards.

Governance and Institutional Framework

CERAI is an expert advisory commission established under the auspices of MIRRI. Membership is multidisciplinary, appointed by the minister with typical term lengths of five years. As an advisory body, CERAI has no direct sanctioning authority; instead its outputs — opinions, guidelines and mappings — inform ministerial policy, public administration practices and legislative proposals. The commission maintains a public-facing secretariat hosted by MIRRI and publishes outputs and contact points on the ministry website; see the founding announcement and member list at MIRRI – appointment announcement. CERAI also curates the national "map of experts" to increase transparency about expertise and to facilitate cross-sectoral engagement: MIRRI – Map of Experts. Institutional links are encouraged with the national Data Protection Authority and sector regulators to align ethical guidance with regulatory obligations.

Key Focus Areas

CERAI concentrates on a set of interlinked priorities: (1) Protecting fundamental rights by ensuring that AI deployments do not unfairly discriminate or disproportionally interfere with privacy and civil liberties; (2) Promoting transparency and explainability, especially for systems used in public administration and safety-critical contexts; (3) Supporting risk management frameworks including impact assessments for high-risk systems; (4) Advising on safety testing, validation and evaluation standards to reduce harms from malfunctioning or misapplied models; (5) Encouraging proactive data protection and privacy-by-design across AI lifecycles; (6) Strengthening accountability and documentation practices, including model documentation and audit trails; (7) Facilitating conformity assessment and potential registration mechanisms for particularly sensitive systems; (8) Addressing cybersecurity and model integrity issues such as robustness to adversarial attacks and supply-chain risk; (9) Scoping sectoral guidance for healthcare and finance where AI decisions can generate significant harm; and (10) Fostering international alignment with EU-level AI regulation and global best practices. These focus areas are reflected in CERAI’s publicly available outputs and expert engagements and are designed to be operationalised through targeted guidance, sectoral recommendations and collaboration with regulators.

Implementation Framework

CERAI implements its mandate through a combination of expert working groups, public consultations, thematic position papers and collaboration with ministerial and sectoral stakeholders. The commission issues non-binding guidance and opinion statements that ministries and public bodies are encouraged to consider in procurement, standard-setting and legislative drafting. The secretariat at MIRRI supports coordination and publication of outputs. CERAI also maintains an expert registry to provide rapid access to ad hoc expertise for ministries and public institutions. For public bodies seeking to deploy AI, CERAI recommends a staged governance approach: risk categorisation; mandatory impact assessment for high-risk cases; independent technical and ethical review; and documentation and monitoring post-deployment. CERAI’s materials include templates and checklists for public administration use and have been referenced in sectoral guidance and academic outreach materials; the main repository of materials and announcements is at MIRRI – CERAI.

Monitoring and Evaluation

CERAI promotes continuous monitoring and evaluation practices focused on system performance, fairness metrics, privacy incidents and emergent harms. Monitoring recommendations include regular re-evaluation of model inputs and outputs, logged evidence of decision pathways for auditability, and post-deployment impact reviews at defined intervals. For public sector uses, CERAI advises that monitoring results be reported to the responsible ministry with summaries available to oversight bodies. Where appropriate, the commission recommends independent audits by qualified third parties and publication of non-sensitive aggregate monitoring outcomes. CERAI also advocates for measurable KPIs in public AI projects (accuracy, false positive/negative rates by subgroup, robustness tests) to enable evidence-based adjustment and accountability.

Penalties, Liability, and Appeals

As an advisory body, CERAI does not itself impose penalties. The commission’s work clarifies pathways for liability and redress by recommending alignment between ethical guidance and existing legal liability frameworks, including consumer protection and administrative law. CERAI’s outputs emphasise the need for clear attribution of responsibility in procurement contracts and operational governance (e.g., data controllers/processors roles under the GDPR). Where legal non-compliance is identified, CERAI refers matters to competent regulators — for example, data protection issues to the national Data Protection Authority — who may exercise enforcement powers under applicable laws. The commission also recommends accessible remedies for individuals harmed by AI decisions and clear internal appeal procedures within public bodies that deploy algorithmic systems.

Relationship to Other Instruments

CERAI’s guidance is designed to complement existing Slovak and EU instruments (e.g., GDPR, national administrative law, sectoral regulations) rather than to supersede them. The commission contributes to national policy formation and aligns recommended practices with EU initiatives on trustworthy AI. CERAI outputs are often referenced by ministries and public stakeholders to ensure coherence between ethical guidance, procurement rules and regulatory obligations. The commission therefore serves a bridging function: translating high-level ethical principles into operational guidance compatible with both national legal frameworks and emerging EU regulation.

International Alignment

CERAI explicitly seeks alignment with EU and international norms and frameworks. In its advisory role the commission tracks developments at the EU level and offers positions intended to harmonise Slovak practices with the European approach to AI governance. CERAI engages academic, civil society and international partners to incorporate comparative regulatory experience, and it encourages Slovak public bodies to adopt assessment and documentation practices consistent with EU proposals on AI risk classifications, transparency and conformity assessment. The commission’s emphasis on interoperability and shared standards is intended to reduce fragmentation and support cross-border legal certainty.

Implementation Timeline

Sources and References