The Duty-Primacy Test: Measuring the Operational Completeness of Algorithmic Transparency Registers

As public-sector algorithmic registries proliferate globally, contemporary critique primarily focuses on baseline compliance deficits (under-reporting) or purely technical opacity (lack of data lineage). This paper introduces a distinct, public-law diagnostic methodology: the Duty-Primacy Test. The framework evaluates algorithm registers not by what they disclose about an automated system’s underlying technology, but by whether the disclosure structurally permits an affected citizen to reconstruct a viable, operational legal pathway to an administrative remedy. We argue that an administrative right is only operationally complete when a vulnerable citizen can explicitly chart three distinct dimensions directly within a public record: a specifically accountable Senior Responsible Owner (SRO) or oversight office, an ascertainable breach mechanism governed by clear evidentiary standards and review procedures, and a public-law remedy capable of enforcing individual rectification or mandatory model suspension. This paper outlines the theoretical architecture, core structural hypotheses, and the rigid qualitative coding protocol—anchored by the "Express Disclosure Principle"—applied to a comprehensive census of the United Kingdom’s Algorithmic Transparency Recording Standard (ATRS) corpus. By shifting the evaluation framework from descriptive transparency to operational enforceability, this methodology exposes the "transparency-operability gap," distinguishing superficial public disclosure from true institutional accountability. Change Log: Version 1 to Version 2 This log tracks the specific refinements made to transition the abstract from a baseline transparency discussion into a doctrinally precise public-law framework: Elevated Target Accountability Roles (The Duty Dimension): Upgraded generic identifiers ("assigned duty-bearers," "individual or office") to "Senior Responsible Owner (SRO) or oversight office." This anchors the abstract in established public-sector governance terminology, particularly within the UK administrative state. Hardened Harm Metrics (The Breach Dimension): Expanded the criteria for identifying breaches from generic "ascertainable breach standards" to "an ascertainable breach mechanism governed by clear evidentiary standards and review procedures." This establishes a measurable legal-doctrinal diagnostic rather than a vague technical threshold. Specified Legal Remedies (The Remedy Dimension): Replaced "effective remedies" and general "systemic rectification" with the explicit terms "public-law remedy" and "mandatory model suspension." This defines a concrete, actionable legal and technical endpoint for holding automated systems accountable. Reframed the Subject (Normative Stakes): Changed the phrase "affected citizen" to "vulnerable citizen" in the core definition of operational completeness. This strategically highlights the institutional power asymmetry between the state and the individual, strengthening the normative justification for the Duty-Primacy Test. Harmonized Structural Terminology: Replaced "an affected party" with "a vulnerable citizen" to maintain stylistic and conceptual continuity with the updated Express Disclosure Principle. Updated Scoring Rubric This updated rubric translates the advanced public-law diagnostics of Version 2 into concrete, reproducible scoring criteria for coders. Domain 0 (Failed) 1 (Partially Satisfied) 2 (Substantially Satisfied) Duty Absent: No role, individual, or department is designated within the record. Monolithic/Implied: The entry identifies the deploying public body as an institutional monolith, or names a generic corporate entity/department without defining specific scope boundaries or legal oversight roles. Explicit SRO: The record explicitly names a specifically accountable role, office, or Senior Responsible Owner (SRO) tasked with legal oversight and system-wide corrective action, alongside an explicitly defined scope of oversight. Breach Absent: No mention of system error, failure rates, review mechanisms, or criteria for challenging outputs. Generalized/Inferable: The record asserts the existence of generalized human review ("human-in-the-loop") or standard internal logs, but fails to provide clear diagnostic thresholds or procedures to evaluate tool errors. Explicit Standards: The entry explicitly specifies the exact evidentiary standards, review procedures, or criteria required for a citizen to demonstrate that an algorithmic output defectively infected a public-law outcome. Remedy Absent: No channels for appeal, complaints, or feedback are provided. Generic/Corporate: The record routes users to generic corporate complaint channels, general contact emails, or standard institutional feedback loops with no fixed timelines or enforceable administrative power. Explicit Public-Law Remedy: The record charts a designated legal or administrative route capable of providing timely individual rectification, initiating an official appeal, or enforcing a mandatory model suspension or system rollback. Key Logic Enforced by the Update • Elimination of Coder Subjectivity: Under the Express Disclosure Principle, if a record mentions a general feedback form but doesn't explicitly guarantee a public-law pathway or model suspension mechanism, coders are now strictly bound to score it a 1 (Partially Satisfied) rather than guessing its effectiveness. • The "Monolith" Filter: In Domain 1 (Duty), naming an entire government agency (e.g., "The Department for Work and Pensions") automatically caps the score at a 1. A score of 2 strictly requires an individual office or SRO designation. • Doctrinal Alignment: The breach and remedy definitions are now anchored (for now, may change later) #dutyprimacy