Computational Trust: Reframing Entity Authority as Annotation Efficiency in AI-Mediated Information Retrieval

This paper argues that algorithmic trust is not a reputation signal but a computational efficiency property: the measurable ease with which a system can evaluate new claims about an entity against its existing entity graph. Computational Trust is formalised as the inverse of marginal annotation cost - when new information fits the existing entity graph cheaply, the system processes it with high confidence and low latency; when it does not fit, the system hedges, re-evaluates, or rejects. The paper introduces the Entity Context Model (the cached entity-specific knowledge structure against which new claims are evaluated), the Annotation Cost Differential (the measurable expression of trust between competing entities), the Confidence Threshold (the binary gate separating low-trust from high-trust processing), and Corroboration Decay (the structural erosion of entity graph confidence through third-party content deletion). Grounded in practitioner evidence from 73 million brand profiles tracked since 2015 and the 2024 Google documentation disclosure confirming binary entity-URL attributes. Part of The Kalicube Process formal foundation.

AI usage disclosure: in the preparation of this paper, the author used Claude (Anthropic) for drafting iterations, structural review, terminology consistency checking, and copy-editing. All conceptual content, theoretical claims, and intellectual contributions are the author's own. The author reviewed, revised, and approved all text and takes full responsibility for the content of this publication.

This paper is TKF-3-18735061 in The Kalicube Framework series by Jason Barnard (Kalicube). Canonical citation identifiers for the whole research programme are maintained in The Kalicube Framework Programme Register (TKF-0): https://doi.org/10.5281/zenodo.20645889. Cite this paper as TKF-3-18735061 (Zenodo concept DOI 10.5281/zenodo.18735061).