Reconstructing Administrative Due Process in Algorithm-Based Government Decision-Making

Abstract

This study examines the adequacy of Indonesia's administrative law framework in safeguarding due process when government decisions are made or materially influenced by algorithmic systems. Using prescriptive doctrinal legal research, the study integrates statutory, conceptual, case, and functional comparative approaches and maps the legal framework through 54 indicators within a Six-Layer Administrative Due Process Model (DPA-6). The mapping identifies 19 indicators as available, 24 as ambiguous, and 11 as not yet specifically regulated. Existing law provides important foundations through legality, the General Principles of Good Governance, personal data protection, administrative remedies, electronic evidence, audit mechanisms, the Ombudsman, and administrative adjudication. The principal gaps concern meaningful human review, risk classification, algorithmic impact assessment, bias testing, system versioning, contestable explanations, and decision-level audit trails. The study proposes DPA-6 as an integrated lifecycle model covering legality and risk, pre-decision duties, notice and explanation, meaningful human review, effective administrative remedies, and audit, adjudication, and remediation. The model links citizens' rights, official responsibility, documentation, oversight, and remedies while preserving proportionality according to decision risk.