04Governing the Meaning Infrastructure Design Implications of Translation Drift in AI- and Artefact-Mediated Governance

Programme Context This preprint forms part of a research programme examining decision systems as longitudinal interpretive-learning architectures. The programme develops a coherent theoretical pipeline linking (1) formal decision-learning architecture, (2) translation drift as a structural mechanism of interpretive misalignment, (3) methodological pathways for making translation coherence empirically observable, and (4) design implications for governance in AI- and artefact-mediated environments. Together, the papers treat governance infrastructures as meaning infrastructures and position institutional learning as the maintenance of interpretive coherence over time. Preprint Description This conceptual, design-oriented article examines the implications of translation drift for governance in environments increasingly mediated by artefacts and AI systems. It argues that contemporary governance infrastructures — templates, scoring models, dashboards, metrics, and algorithmic systems — function not only as decision supports but as meaning infrastructures that stabilise and transmit institutional interpretations of value, risk, feasibility, and success. Building on prior work identifying translation drift as a structural mechanism through which interpretive coherence can decay across layered governance artefacts, the article shows how digital mediation relocates drift into infrastructure. Evaluative assumptions may become embedded in criteria structures, model architectures, optimisation logics, and dashboards, where they can persist and propagate across decision cycles. Under these conditions, drift may arise not only through gradual human reinterpretation but through the infrastructural hardening of historical meaning. The paper does not propose specific technical controls or optimisation methods. Instead, it develops design-oriented implications for governance systems. Governance is reframed as interpretive maintenance — the ongoing work of keeping criteria, models, metrics, and review practices aligned in how they express institutional intent. From this perspective, organisations require capabilities to inspect their own decision logic, exercise meta-governance over artefact ecosystems, and position AI systems as instruments of interpretive traceability rather than autonomous decision authorities. The analysis is relevant for organisational learning, governance design, digital portfolio management, AI governance, and institutional decision systems operating under conditions of scale, complexity, and delayed feedback. Version 2.00 This is the first public release of this manuscript within the research programme structure. The paper presents the design-oriented analysis linking translation drift to meaning infrastructure and governance implications. Cross-paper terminology has been harmonised, the unit-of-analysis statement has been standardised across the series, and reference architecture has been aligned. The positioning of AI as interpretive traceability support is consistent with the architectural and methodological papers. No empirical datasets are associated with this version.