Algorithm in the Structure of Administrative Decision-Making: A Model of Explainability and Burden of Proof for an Effective Legal Remedy within Discretionary Power in Public Administration
Open accessThe DOI matches its Crossref record; none of the 25 sources its claims cite is checked for support yet.
Abstract
The introduction of artificial intelligence (AI) systems into public administration decision-making processes poses fundamental challenges to procedural guarantees, particularly the right to an effective legal remedy. This article examines how the use of algorithms, especially those of a “black box” nature, affects the transparency of proceedings and the possibility of judicial review. Based on a dogmatic analysis, rooted in the concept of a rationalized administrative decision, and a comparative case study analysis (Poland, the Netherlands, Estonia, Finland), the author argues that the lack of explainability in AI systems paralyzes the right to appeal. In response to the diagnosed problems, the article proposes a model for a minimum standard of “algorithmic justification”, which can be implemented within the legal system. This model aims to restore transparency, reverse the unfavorable burden of proof for the individual, and adapt the judicial cognition model to the new technological reality, taking into account ESG frameworks as a standard of due diligence for public authorities.
Overview
CoChat Summary
TL;DR
The paper argues that opaque "black box" AI systems in public administration undermine the right to an effective legal remedy by making administrative decisions unverifiable and shifting an impossible burden of proof onto individuals. It proposes a minimum standard of "algorithmic justification" — centered on an "Algorithm Card" and a "Transparency and Explainability Checklist" — embeddable in the Polish Administrative Procedure Code and supported by the GDPR, AI Act, and ESG/DEIB frameworks.
Key findings
- The lack of explainability in AI systems "paralyzes the right to appeal": without access to the algorithm's logic, neither the individual nor the administrative court can reconstruct the authority's reasoning, making judicial review of legality "illusory" and contradicting the principle of persuasion (Article 11 APC) and the principle of deepening citizens' trust (Article 8 APC).
- The Dutch childcare benefits scandal (*Toeslagenaffaire*) is presented as confirmation of hypothesis H1: an opaque, self-learning risk-profiling algorithm "automatically and often discriminatorily flagged thousands of families as potential fraudsters," producing a cascade of erroneous decisions and financial ruin — the antithesis of the rationalized decision model.
- The proposed "Algorithm Card" (inspired by Model Cards for Model Reporting) should document at minimum: purpose and scope of application, input data description, general outline of decision-making logic, system version number, and basic quality/performance metrics (e.g. accuracy, error rate).
- The model reverses the burden of proof onto the authority: the more automated and complex the process, the higher the standard of proof and the more complete the justification required for the decision to count as rationalized and reviewable.
Methods
- Design: integrated legal analysis combining dogmatic analysis (using the concepts of the rationalized administrative decision and discretionary space) with a comparative case study analysis, culminating in a synthesis for *de lege ferenda* postulates.
- Sample/cases: administrative practices in Poland (2023 Supreme Audit Office report on advanced analytical tools), the Netherlands (childcare benefits scandal; Algorithm Register), Estonia (KrattAI strategy), and Finland (AuroraAI ethical principles).
- Analysis: confrontation of practice with the theoretical model to identify universal risks (e.g. erosion of the right to appeal) and best practices (e.g. proactive governance frameworks); three working hypotheses (H1–H3) tested.
Limitations
- Stated by the paper: the subject scope is limited to administrative law within the Polish and EU legal orders, excluding the private sector; the analysis is legal in nature and "intentionally omits the deeply technical aspects of AI system development," focusing on procedural and substantive legal consequences.
- Evident from the text: the case studies rely on secondary sources (audit reports, government strategies, a parliamentary inquiry) rather than original empirical data, and the proposed model is a normative proposal not yet tested in practice.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 5 Not valid · 23 Not checked · 49 Own results and interpretations
In the paper
“Early practice already shows that this risk is not merely theoretical but can lead to real violations of civil rights on a massive scale.”
What the sources say
- Not matchedcites [4]
The reference list we hold has no entry [4].
plainstates a cause- 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 conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The Dutch Algorithm Register can be interpreted as a practical attempt to re -store the conditions necessary for making rationalized decisions in the digital era.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: unclear — It is unclear whether it could be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The introduction of artificial intelligence (AI) systems into public administration decision-making processes poses fundamental challenges to procedural guarantees, particularly the right to an effective legal remedy.”
plainstates a cause- 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.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The introduction of AI into this process does not eliminate discretionary space but rather shifts part of it into the realm of algorithm design and operation.”
plainstates a cause- 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.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In this way, the EU legal framework directly supports the theoretical postulate of verifiability, which is the foundation of a rationalized administrative decision.”
plainstates a cause- 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.
15 more claims
Create a free account to see all 77 claims and every source verdict.
- Probably not a substantial claim. 43
- It may state two separate claims as one. 26
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
§ INTRODUCTION · p. 8
“25 M. Mitchell, S. Wu, N. Gessner, A. Zaldivar, P. Barnes, A. Vasserman, B. Hutchinson, E. Spitzer, I.D. Raji, T. Gebru, Model Cards for Model Reporting”
§ DISCUSSION AND DE LEGE FERENDA POSTULATES · p. 13
“Mitchell M., Wu S., Zaldivar A., Barnes P., Vasserman A., Hutchinson B., Spitzer E., Raji I.D., Gebru T., Model Cards for Model Reporting”
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 1
- Parliamentary Inquiry into Child Benefit, Ongekend Onrecht (Unprecedented Injustice), 2020, https:// www.tweedekamer.nl/sites/default/files/atoms/files/20201217\_eindverslag\_parlementaire\_on -dragingscommissie\_kinderopvangtoeslag.pdf (access: 29.12.2025). UK Information Commissioner's Office (ICO), Guidance on AI and Data Protection , https://ico.org. uk/for-organisations/uk-gdpr-guidance-and-resources/artificial-intelligence/guidance-on-ai-anddata-protection (access: 29.12.2025). Legal acts Act of 14 June 1960 - Administrative Procedure Code (consolidated text, Journal of Laws 2025, item 1691, as amended). Constitution of the Republic of Poland of 2 April 1997 (Journal of Laws 1997, no. 78, item 483, as amended). Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) (OJ L 119/1, 4.5.2016). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No. 300/2008, (EU) No. 167/2013, (EU) No. 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) (OJ L 2024/1689, 12.7.2024). ABSTRAKT UMCS
Metadata
- Authors
- Szot, A.
- Affiliations
- Maria Curie-Skłodowska University (PL)
- Corresponding author
- Szot, Adam
- ORCID
- Szot, Adam
- Venue
- Studia Iuridica Lublinensia
- Publisher
- Maria Curie-Sklodowska University
- Published
- 2026-01-07
- Volume
- 34
- Issue
- 5
- Pages
- 269-282
- ISSN
- 1731-6375 (ISSN-L)
- 2449-8289
- OpenAlex
- W7119011567
- Cited by
- 3
- References
- 0
- Citations by year
- 2026: 3
- FWCI
- 59.23
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- cc-by
- Locations
- Studia Iuridica Lublinensia (publishedVersion, cc-by)
- Studia Iuridica Lublinensia (Uniwersytet Marii Curie-Skłodowskiej w Lublinie) (publishedVersion, cc-by)
- Topics
- Social Sciences › Law › Administrative Law and Governance
- Social Sciences › Law › Digitalization, Law, and Regulation
- Social Sciences › Law › Artificial Intelligence in Law
- Keywords
- algorithmic decision-making, artificial intelligence in public administration, administrative discretion, burden of proof, algorithmic explainability, black-box algorithms, judicial review, public sector organizations, procedural guarantees
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.
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).
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: Algorithm in the Structure of Administrative Decision-Making: A Model of Explainability and Burden of Proof for an Effective Legal Remedy within Discretionary Power in Public Administration. https://app.cochat.ai/papers/10.17951/sil.2025.34.5.269-282
- The paper
- Szot, A. (2026). Algorithm in the Structure of Administrative Decision-Making: A Model of Explainability and Burden of Proof for an Effective Legal Remedy within Discretionary Power in Public Administration. Studia Iuridica Lublinensia. https://doi.org/10.17951/sil.2025.34.5.269-282
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.