Legal Accountability in Algorithmic Governance and Public Decision-Making in Administrative Law

The article begins by explaining that the rapid expansion of algorithmic systems in public administration has changed how public decisions are produced, processed, and justified. Government institutions increasingly use data-driven systems for classification, prediction, prioritization, and evaluation. These systems introduce a new layer of decision-making that operates through technical procedures rather than direct human reasoning alone.

The article emphasizes that algorithmic tools are becoming embedded in administrative routines and therefore influence the structure of authority within public institutions. Public decision-making is no longer shaped only by legal rules, bureaucratic discretion, and institutional procedures. It is increasingly mediated by computational systems that influence how facts are interpreted and how outcomes are generated.

The central problem is that legal accountability frameworks have not developed at the same pace as algorithmic governance. Administrative law traditionally assumes that public decisions can be traced to identifiable actors who can justify and defend their actions. This assumption becomes unstable when decisions are shaped by opaque or semi-autonomous algorithmic systems.

The article argues that affected individuals may face public decisions that are difficult to understand, challenge, or review through ordinary legal mechanisms. Algorithmic systems can affect access to public services, administrative status, and governmental treatment. When responsibility becomes blurred, the legitimacy of public authority also becomes uncertain.

Existing scholarship has already recognized that algorithmic governance raises normative and institutional concerns. Many studies focus on fairness, bias, transparency, explainability, regulatory compliance, and procedural safeguards. These discussions show that algorithmic systems create risks for rights, equality, and due process.

However, the article identifies a legal-conceptual gap. Existing debates often treat ethics, compliance, and accountability as separate issues rather than parts of one unified legal problem. There is still limited clarity about how law should reconstruct responsibility when decision-making is distributed across humans, institutions, and systems.

The article argues that the research gap lies in the disconnect between technological transformation and legal reconstruction. Public administration is increasingly shaped by algorithmic systems that alter responsibility, discretion, and evidence production, but legal analysis has not yet fully developed a coherent framework for accountability under these conditions.

The article is guided by questions about how algorithmic governance challenges conventional legal accountability, why traditional responsibility and review frameworks are inadequate, how algorithmic accountability should be understood as a legal obligation, and what elements are necessary for reconstructing accountability. It proposes transparency, contestability, and institutional oversight as core elements for preserving lawful public administration under algorithmic conditions.

The article uses a qualitative research design with a normative-conceptual analytical framework. This method is appropriate because the study examines the meaning, structure, and transformation of legal responsibility rather than measurable behavior or causal effects. The research focuses on how accountability is conceptually reconstructed when administrative decisions are shaped by algorithmic systems that complicate responsibility, reviewability, and justification.

The analytical framework combines administrative law, legal theory, and governance analysis. The study uses close reading, categorization of core concepts, and systematic extraction of arguments related to responsibility, reviewability, justification, opacity, and distributed decision-making. A concept matrix is used to organize key legal dimensions, including transparency, contestability, institutional oversight, evidentiary opacity, distributed responsibility, and the boundary between human discretion and automated judgment.

The article finds that algorithmic governance disrupts the legal structure of public administration because it changes how authority is exercised. Conventional accountability assumes that a decision can be traced to a legally recognized official who can explain the reasons, defend the evidence, and submit the decision to review. This model becomes unstable when administrative outcomes are shaped by opaque algorithmic systems.

The main legal problem is not only technological opacity. It is the weakening of the link between public power and answerable decision-making. When algorithms classify, rank, predict, or recommend outcomes, affected persons and even public officials may not fully understand how the decision was produced.

The article argues that existing accountability frameworks are inadequate because they were built for human-centered administration. These frameworks are less capable of addressing distributed responsibility, system-mediated judgment, and opaque justification. Algorithmic governance may preserve the appearance of legality while altering the conditions that make legality meaningful.

In many administrative contexts, officials still issue the final decision, but the substance of the decision is shaped by automated scoring, pattern recognition, or algorithmic classification. Formal human involvement does not automatically preserve accountability if the operative reasoning has already been delegated to a system.

Responsibility becomes fragmented across officials, software developers, external vendors, data managers, and supervisory bodies. Each actor may contribute to the decision process, but none may fully own the final outcome. This makes legal attribution difficult even though the decision still affects rights, benefits, status, or public treatment.

The article criticizes fragmented responses that focus only on ethics, fairness, bias, explainability, compliance, or documentation. These responses are useful but incomplete. Ethics may identify desirable norms, and compliance may create formal controls, but legal accountability requires institutional duties, review structures, and standards of justification.

The article proposes that algorithmic accountability should be understood as a composite legal obligation. Transparency is necessary so that affected individuals and reviewing bodies can understand how a decision was shaped. However, transparency alone is not enough if the disclosed information is too technical or inaccessible to support a legal challenge.

Contestability is also necessary because affected individuals must be able to question, correct, and review algorithmically shaped decisions. Without contestability, rights of review become merely formal. Effective contestability requires accessible appeal procedures, reason-giving duties, and mechanisms for human reconsideration.

Institutional oversight completes the accountability framework. Oversight ensures that accountability is not reduced to private explanation or individual complaint. Public mechanisms of supervision, audit, intervention, and correction are needed to maintain control over algorithmic systems used in public decision-making.

The article explains that algorithmic systems change the meaning of administrative discretion. In conventional administration, discretion involves legally structured human judgment. Under algorithmic governance, discretion may be narrowed, displaced, or shaped by system design, thresholds, classifications, and automated recommendations.

Evidentiary opacity is another major challenge. Legal review normally requires facts and reasons that can be examined by affected persons and reviewing bodies. Algorithmic outputs may be based on data correlations and processing logic that are difficult to translate into reviewable evidence, weakening the legal visibility of administrative decisions.

The article concludes that public law must treat algorithmic systems as part of the administrative decision structure, not merely external technical support. Accountability must remain attached to identifiable public authority even when decision-making is distributed. Public institutions must explain the operational role of algorithmic systems, preserve evidentiary reviewability, and ensure meaningful human judgment where legally significant decisions are made.

The expansion of algorithmic governance has altered the legal foundations of public decision-making by introducing opaque, distributed, and system-mediated forms of administrative judgment. Conventional accountability frameworks are increasingly insufficient because they were developed for human-centered administrative action, where responsibility, justification, and reviewability could be more clearly located within identifiable institutional actors. The discussion established that algorithmically shaped decisions complicate these assumptions by diffusing responsibility across technical and administrative arrangements, weakening evidentiary visibility, and narrowing the practical space for meaningful legal review. Under these conditions, accountability can no longer be understood as a single procedural requirement. It must be reconstructed as a composite legal obligation grounded in transparency, contestability, and institutional oversight so that public authority remains legally answerable even when decision processes are shaped by automated systems.

The main contribution lies in advancing a more integrated legal-conceptual framework for understanding accountability in algorithmic public administration. By bringing administrative law, legal theory, and governance analysis into one analytical structure, the discussion clarifies that the core challenge is not only the presence of new technology, but the transformation of the institutional conditions through which legality is maintained. This perspective moves beyond fragmented debates that treat ethics, compliance, and accountability as separate domains, and instead positions algorithmic accountability within the internal architecture of lawful administration. The framework strengthens current scholarship by explaining how transparency, contestability, and oversight function as interdependent legal dimensions rather than isolated regulatory tools. In doing so, it provides a clearer basis for analyzing how public decision-making can remain legitimate when authority is increasingly exercised through socio-technical systems.

Future research needs to extend this framework into more specific administrative settings in order to examine how legal accountability operates across different forms of algorithmic decision-making. Greater attention is needed to the institutional design of review mechanisms, especially in situations where administrative officials rely heavily on automated outputs that are difficult to translate into legally intelligible reasons. Further inquiry is also necessary on the allocation of responsibility across public agencies, private vendors, and technical infrastructures, particularly where authority is formally public but operationally distributed. Comparative legal studies would be valuable for assessing how different administrative systems respond to opacity, evidentiary complexity, and the changing boundary between human discretion and automated judgment. Deeper doctrinal and governance-oriented research in these areas would help refine the legal standards required to preserve accountability as algorithmic systems become more deeply embedded in public administration.

Ananny, M., & Crawford, K. (2018). Seeing without knowing: Limitations of the transparency ideal and its application to algorithmic accountability. New Media & Society, 20, 973–989. https://doi.org/10.1177/1461444816676645

Bateman, W. (2019). Algorithmic Decision-Making and Legality: Public Law Dimensions. https://consensus.app/papers/algorithmic-decisionmaking-and-legality-public-law-bateman/3b8feab24c265eea850c39cac7739e43/

Baykurt, B. (2022). Algorithmic accountability in U.S. cities: Transparency, impact, and political economy. Big Data & Society, 9. https://doi.org/10.1177/20539517221115426

Binns, R. (2017). Algorithmic Accountability and Public Reason. Philosophy & Technology, 31, 543–556. https://doi.org/10.1007/s13347-017-0263-5

Brauneis, R., & Goodman, E. (2017). Algorithmic Transparency for the Smart City. LSN: Rights & Liberties (Topic). https://doi.org/10.2139/ssrn.3012499

Butler, O. (2025). Algorithmic Decision-Making, Delegation and the Modern Machinery of Government. Oxford Journal of Legal Studies, 45, 727–752. https://doi.org/10.1093/ojls/gqaf018

Cohen, T., & Suzor, N. (2024). Contesting the public interest in AI governance. Internet Policy Review, 13. https://doi.org/10.14763/2024.3.1794

Cole, R. (2023). Inter-Rater Reliability Methods in Qualitative Case Study Research. Sociological Methods & Research, 53, 1944–1975. https://doi.org/10.1177/00491241231156971

Elo, S., Kääriäinen, M., Kanste, O., Pölkki, T., Utriainen, K., & Kyngäs, H. (2014). Qualitative Content Analysis. SAGE Open, 4. https://doi.org/10.1177/2158244014522633

Enarsson, T., Enqvist, L., & Naarttijärvi, M. (2021). Approaching the human in the loop – legal perspectives on hybrid human/algorithmic decision-making in three contexts. Information & Communications Technology Law, 31, 123–153. https://doi.org/10.1080/13600834.2021.1958860

Engelmann, A. (2023). Algorithmic transparency as a fundamental right in the democratic rule of law. Brazilian Journal of Law, Technology and Innovation. https://doi.org/10.59224/bjlti.v1i2.169-188

Feng, N., & Chandra, Y. (2025). Accountability in Government Use of AI: Citizen Concerns and Preferences. Public Administration. https://doi.org/10.1111/padm.70030

Gules-Guctas, E. (2025). How Do Algorithmic Decision-Making Systems Used in Public Benefits Determinations Fail? Insights From Legal Challenges. Public Administration Review. https://doi.org/10.1111/puar.70043

Kemper, J., & Kolkman, D. (2018). Transparent to whom? No algorithmic accountability without a critical audience. Information, Communication & Society, 22, 2081–2096. https://doi.org/10.1080/1369118x.2018.1477967

Langer, C. (2024). Decision-making power and responsibility in an automated administration. Discover Artificial Intelligence, 4. https://doi.org/10.1007/s44163-024-00152-1

Linneberg, M., & Korsgaard, S. (2019). Coding qualitative data: a synthesis guiding the novice. Qualitative Research Journal. https://doi.org/10.1108/qrj-12-2018-0012

Liu, H.-W., Lin, C.-F., & Chen, Y.-J. (2019). Beyond State v Loomis: artificial intelligence, government algorithmization and accountability. International Journal of Law and Information Technology, 27, 122–141. https://doi.org/10.1093/ijlit/eaz001

Matteucci, S. (2021). Public Administration Algorithm Decision-Making and the Rule of Law. European Public Law. https://doi.org/10.54648/euro2021005

Oswald, M. (2018). Algorithm-assisted decision-making in the public sector: framing the issues using administrative law rules governing discretionary power. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 376. https://doi.org/10.1098/rsta.2017.0359

Panagopoulou, F. (2024). Algorithmic Decision-Making in Public Administration. Journal of Public Administration. https://doi.org/10.22259/2642-8318.0601001

Rieder, B., & Hofmann, J. (2020). Towards platform observability. Internet Policy Review, 9. https://doi.org/10.14763/2020.4.1535

Saldanha, D. M. F., Dias, C., & Guillaumon, S. (2022). Transparency and accountability in digital public services: Learning from the Brazilian cases. Government Information Quarterly, 39, 101680. https://doi.org/10.1016/j.giq.2022.101680

Scassa, T. (2020). Administrative Law and the Governance of Automated Decision-Making: A Critical Look at Canada’s Directive on Automated Decision-Making. Social Science Research Network. https://doi.org/10.2139/ssrn.3722192

Suksi, M. (2020). Administrative due process when using automated decision-making in public administration: some notes from a Finnish perspective. Artificial Intelligence and Law, 29, 87–110. https://doi.org/10.1007/s10506-020-09269-x

Tsarouhas, P., & Grigoriadis, K. (2025). Building Trust in AI for Public Administration: A Strategic Framework for Transparency, XAI, Participation, and Digital Literacy. 2025 7th International Congress on Human-Computer Interaction, Optimization and Robotic Applications, 1–9. https://doi.org/10.1109/ichora65333.2025.11017116