AUTOMATED DECISION-MAKING IN ADMINISTRATIVE PENAL MATTERS
Open accessThe DOI matches its Crossref record; 5 of the 5 references it cites were found in a registry or an index; 1 of the 39 sources its claims cite were checked for support.
Abstract
Automated decision-making in administrative punishment brings new challenges in the area of legal regulation and state responsibility. This article analyses the existing legal regulations that determine the conditions of liability for damage caused by artificial intelligence decisions in administrative law, with an emphasis on administrative punishment. Particular attention is paid to Act No. 514/2003 Coll. on liability for damage caused in the exercise of public authority and its applicability to cases of incorrect official procedures of systems using artificial intelligence. The paper also assesses the impact of European legislation, in particular the Artificial Intelligence Act (AIA) and the GDPR, on liability relationships in public administration. The article also deals with mechanisms for judicial review of artificial intelligence decisions, the need for transparency and ethical issues of automated decisionmaking. It concludes by identifying key challenges and recommendations for the regulation of artificial intelligence in administrative punishment.
Overview
CoChat Summary
TL;DR
The paper analyses whether and under what conditions automated decision-making may lawfully be used in Slovak administrative punishment, focusing on state liability under Act No. 514/2003 Coll. and the impact of the GDPR and the EU Artificial Intelligence Act. It concludes that automation is permissible only where rule-of-law principles (legality, transparency, proportionality, defence rights, effective remedies) are upheld, and that "hybrid" decision-making dilutes responsibility.
Key findings
- Automated decision-making in administrative punishment is lawful only if four conditions are met: a clear legal basis for the authority's use of automated tools; processing of accurate, up-to-date, lawfully processable data; transparency toward data subjects (notification plus explanation of automation's role); and respect for fundamental rights, especially the right to a fair trial.
- The paper argues AI systems do not meet the requirements of legal subjectivity — they lack independent will, intent, consciousness, property and capacity to be sanctioned — so granting them legal personality would create a "legally empty concept"; a "spectral model" of legal subjectivity is discussed but responsibility must remain specific, identifiable and enforceable.
- Act No. 514/2003 Coll. (liability for damage caused in the exercise of public authority), grounded in Article 46(3) of the Slovak Constitution, can provide a basis for liability for unlawful decisions or incorrect official procedure, but so-called hybrid decision-making — where the system formally only "proposes" and the human adopts it automatically — dilutes responsibility and weakens identification of the liable entity.
- The paper proposes a proportionality test before deploying automated systems (criteria: severity of interference, transparency, category of person concerned, existing control mechanisms), a mandatory ex ante impact assessment, algorithmic auditability, the right to individual justification, and a possible shift of the burden of proof toward the public administration given citizens' knowledge asymmetry; it also proposes court-appointed expert explanation of algorithmic decisions.
Methods
- Design: analytical-deductive legal interpretation, supported by comparison of the legal regulation of the Slovak Republic, the Czech Republic and relevant EU law (GDPR, Regulation (EU) 2024/1689 — the AIA), with practical scenarios of AI use in decision-making as a starting point.
- The analysis is structured around three research questions: (1) conditions under which an algorithmic system may decide in administrative punishment without human intervention; (2) who is liable for damage caused by an automated decision and under what circumstances; (3) whether the existing Slovak and EU framework sufficiently protects data subjects' fundamental rights.
- No empirical dataset, sample or statistical analysis is reported; the paper is doctrinal and normative, drawing on case law (e.g. Malige v. France, Jussila v. Finland, Constitutional Court of the Slovak Republic PL. ÚS 10/2014 and I. ÚS 258/2021), Council of Europe recommendations (CM/Rec(2007)7, R (91) 1) and academic literature.
Limitations
- The paper states that in the Slovak Republic the topic "has not yet been systematically addressed by legal theory," with discussion occurring mainly at the ethical-technological level or in connection with digitalisation as a whole; it positions itself as building on Czech research impulses and adding a Slovak perspective.
- Evident from the text: the analysis is purely doctrinal with no empirical testing, case data or quantitative evidence; the liability conclusions are framed as open questions and de lege ferenda proposals rather than settled answers.
- Evident from the text: the paper repeatedly uses hypothetical illustrations (e.g. an automated traffic-offence fine based on incorrectly evaluated data, "for example, in the case of poor calibration of measuring devices") rather than documented instances, so the practical scope of the identified liability gaps is not demonstrated.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
1 Valid · 1 Not valid · 35 Not checked · 27 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This may lead to data subjects avoiding the use of more advanced, albeit more effective, AI technologies in order to comply with the requirements of the GDPR.”
What the sources say
- Partly supportedcites [13] 13. HUBKOVÁ, P. EU Administrative Decision-Making Delegated to Machines - Legal Challenges and Issues. Acta Universitatis Carolinae - Iuridica, 2024, roč. 70, č. 2, s. 22. https://doi.org/10.14712/23366478.2024.25 sourceExperimental — accuracy not yet measured
Section 3.3 (p. 19) says that sophisticated AI can be difficult to explain and that opacity precludes certain self-learning systems under the PNR Directive. Section 3.1 (p. 13) discusses GDPR requirements for automated decisions, but the paper does not say that data subjects avoid more effective AI technologies to comply with the GDPR; it discusses restrictions on the authorities using those tools. (LLM, conf=0.82)
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
- Partly supported
In the paper
“Decisions in administrative penalty cases where automation is used must meet the same requirements as any other decision taken by public authorities.”
What the sources say
- Not matchedcites [3] 3. HIGH-LEVEL EXPERT GROUP ON ARTIFICIAL INTELLIGENCE (AI HLEG). Regulatory framework on AI . [online]. European Commission, [cit. 12. januára 2025]. Dostupné z: https://eucrim.eu/news/ai-high-level-expert-group-publishes-ethicschecklist/.
We searched by its title and found no matching paper.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
In the paper
“Article 22 of the GDPR gives data subjects the right not to be subject to decisions based solely on automated processing, including profiling, which produce legal effects or significantly affect them.”
What the sources say
- Not matchedcites Article 22 of the GDPR
This doesn’t read as a citation we can look up.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
In the paper
“Article 46(1) of the Constitution of the Slovak Republic enshrines the fundamental right of individuals to seek judicial or other protection provided by law.”
What the sources say
- Not matchedcites [25]
The reference list we hold has no entry [25].
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
A statement the paper makes itself — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In this context, the principle of legality requires that administrative sanctions may only be imposed on the basis of a legal authorisation.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
21 more claims
Create a free account to see all 64 claims and every source verdict.
- Probably not a substantial claim. 30
- It may state two separate claims as one. 17
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ II. LEGAL FRAMEWORK AND LIMITS OF AUTOMATION
“However, it does not mention artificial intelligence or related concepts such as autonomous systems, intelligent technologies, profiling, automated decision-making, machine learning or big data.”
§ II. LEGAL FRAMEWORK AND LIMITS OF AUTOMATION
“Article 22 of the GDPR gives data subjects the right not to be subject to decisions based solely on automated processing, including profiling, which produce legal effects or significantly affect them.”
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 2
- [12]12. SVÁK, J. In: OROSZ, L., J. SVÁK a kol. Ústava Slovenskej republiky. Komentár. 1. zväzok (základné princípy a ľudské práva). Bratislava: Wolters Kluwer, 2021, 892 s. ISBN 9788057103813
- [15]15. ALMIOTTO, F. When Is a Decision Automated? A Taxonomy for a Fundamental Rights Analysis. Forthcoming in German Law Review, s. 1-27 Dostupné z: https://papers. ssrn.com/sol3/papers.cfm?abstract\_id=4578761 [cit. 12. 04. 2025].
Metadata
- Authors
- Comenius University Bratislava, F. O. L., & Lengyel, A.
- Affiliations
- Comenius University Bratislava (SK)
- Corresponding author
- Lengyel, Alan
- Venue
- STUDIA IURIDICA Cassoviensia
- Published
- 2025-12-19
- Volume
- 13
- Issue
- 3
- ISSN
- 1339-3995 (ISSN-L)
- OpenAlex
- W7118414988
- Cited by
- 0
- References
- 0
- FWCI
- 0.00
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- cc-by
- Locations
- STUDIA IURIDICA Casoviensia (publishedVersion, cc-by)
- Topics
- Social Sciences › Law › Digitalization, Law, and Regulation
- Social Sciences › Law › Administrative Law and Governance
- Social Sciences › Law › Artificial Intelligence in Law
- Keywords
- automated decision-making, administrative sanctions, AI liability, state liability, EU AI Act, GDPR, transparency, public administration, Czech administrative law
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: AUTOMATED DECISION-MAKING IN ADMINISTRATIVE PENAL MATTERS. https://app.cochat.ai/papers/10.33542/sic2025-s-07
- The paper
- Comenius University Bratislava, F. O. L., & Lengyel, A. (2025). AUTOMATED DECISION-MAKING IN ADMINISTRATIVE PENAL MATTERS. STUDIA IURIDICA Cassoviensia. https://doi.org/10.33542/sic2025-s-07
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.