The Calibrated Authority Index™

The Calibrated Authority Index™ is a novel, original dataset: no prior corpus codes the public generative-AI policies of knowledge institutions as revealed verification-economics judgments. Each institution's public AI policy is coded on six dimensions (D1–D6) into a composite Calibrated Authority (CA) score 0–12, plus posture, verification-boundary fit, trust-logic, and twilight markers. Each record carries the institution's verbatim policy quote and source URL; every score traces to a public document. It is the observational, observed-in-the-wild evidence for the trust-formation-divergence thesis: trust forms by divergent mechanisms driven by verification economics — institutions permit AI where its output can be cheaply verified, and reserve for a human the work that cannot be. N=59 institutions; mean Calibrated Authority 9.7/12. Agent-built and agent-maintained. The Index is produced and kept current by an automated coding pipeline that continuously re-reads and re-scores each institution's policy; new and revised records are released at a human-reviewed pace. This deposit is one human-paced public snapshot of a living corpus. Versioned DOIs. This deposit is wired to a GitHub repository via Zenodo's GitHub integration: each tagged GitHub release mints a new versioned DOI under a single concept DOI, so the citable record tracks the living index over time. Released CC-BY-4.0.