Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 68 of the 68 references it cites were found in a registry or an index; 3 of the 45 sources its claims cite were checked for support.

Artificial Intelligence (AI) and Automation in Administrative Procedures: Potentials, Limitations, and Framework Conditions

Open access
Parycek, Peter, Schmid, Verena, Novak, Anna-Sophie · 2023 · Journal of the Knowledge Economy

Abstract

The paper examines AI integration into administrative procedures and emphasizes data quality, transparency, and system limitations.

Overview

CoChat Summary

TL;DR

The paper examines the potentials, limitations, and framework conditions for using AI and automation in administrative procedures, focusing on Austria and Germany as civil law examples. It argues that rule-based systems fit formalized administrative law and offer transparency, while machine-learning systems require sufficient high-quality, unbiased data and raise explainability and bias concerns.

Key findings

  • Rule-based systems are transparent and suitable for automating administrative procedures where law and facts can be formalized into “if A, then B” rules; full automation is conceivable mainly for simply structured, standardized procedures, such as Austria’s application-free family allowance.
  • Machine-learning systems do not apply legal norms directly and require large amounts of data; the paper notes that approximately 500,000 court rulings are issued annually in Germany, but this may still be too low for AI training depending on the task and model, and cites LexMachina as based on training data from over 100,000 cases.
  • Key challenges include limits in formalizing legal language, administrative discretion, data availability and quality, bias (sampling bias, omitted variable bias, discriminatory bias), and transparency/explainability; machine-learning systems are described as black-box systems, whereas rule-based systems retain comprehensibility of results.
  • Framework conditions are proposed at individual, technical, organizational, and societal levels, including measures against automation bias, critical treatment of “human in the loop,” user-centered design, quality assurance, proxy models, counterfactual explanations, result review, risk management, and continuous monitoring.

Methods

  • The study uses a scoping review plus jurisprudential methods, searching international journals and Austrian and German legal literature with predetermined inclusion and exclusion criteria.
  • The authors considered about 130 sources in more detail and then reduced them to the most essential for the analysis.
  • The legal focus is on Austria and Germany as civil law systems under the principle of legality; the paper is a conceptual and literature-based analysis rather than an empirical implementation study.

Limitations

  • The paper states that it does not attempt to review the entire body of law on AI and law, focuses on examples from Austria and Germany, and is written from a continental European perspective, so Anglo-American scenarios require caution.
  • The paper states that data protection issues are not discussed further and that it remains unclear to what extent the quality of learning results differs between real and synthetic register data.
  • Evident from the text: the proposed framework conditions are derived from a literature review and are not empirically tested; the authors call for continuous monitoring and further research.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s13132-023-01433-3

Cites

68 of 68 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

3 Valid · 12 Not valid · 45 Not checked · 95 Own results and interpretations

41 more claims

Create a free account to see all 155 claims and to verify the 2 references still open.

See all claims — free account
  • It may state two separate claims as one. 45
  • Probably not a substantial claim. 84

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 63 references

Research works printed but not linked: 3

  1. Crawford, K., & Schultz, J. (2019). AI systems as state actors. Columbia Law Review, 119(7), 1941–1972.
  2. Lachana, Z., Alexopoulos, C., Loukis, E., & Charaladibis, Y. (2018). Identifying the different generations of Egovernment: an analysis framework. The 12th Mediterranean Conference on Information Systems (MCIS).
  3. von Blumröder, L., & Breiter, A. (2020b). Die Nutzung maschineller Lernsysteme für den Erlass verwaltungsrechtlicher Entscheidungen. dms. · DOI

Cited by

See all 131

Metadata

Authors
Parycek, P., Schmid, V., & Novak, A.-S.
Affiliations
Universität für Weiterbildung Krems (AT)
Corresponding author
Novak, Anna-Sophie
Venue
Journal of the Knowledge Economy
Published
2023-06-20
Volume
15
Issue
2
Pages
8390-8415
ISSN
  • 1868-7865 (ISSN-L)
  • 1868-7873
OpenAlex
W4381436268
Cited by
131
References
63
Citations by year
2026: 40 · 2025: 60 · 2024: 28 · 2023: 3
FWCI
128.79
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of the Knowledge Economy (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
Funders
Danube University Krems University for Continuing Education
Topics
  • Social Sciences › Law › Administrative Law and Governance
  • Social Sciences › Law › Artificial Intelligence in Law
  • Social Sciences › Law › Digitalization, Law, and Regulation
Keywords
artificial intelligence, administrative procedure, administrative law, automation, algorithmic transparency, data quality, algorithmic bias, machine learning, rule-based systems, trust in AI, data availability

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: Artificial Intelligence (AI) and Automation in Administrative Procedures: Potentials, Limitations, and Framework Conditions. https://app.cochat.ai/papers/10.1007/s13132-023-01433-3
The paper
Parycek, P., Schmid, V., & Novak, A.-S. (2023). Artificial Intelligence (AI) and Automation in Administrative Procedures: Potentials, Limitations, and Framework Conditions. Journal of the Knowledge Economy. https://doi.org/10.1007/s13132-023-01433-3

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.