AIUC-1 Explorer

Methodology

AIUC-1 Explorer takes successive releases of the AIUC-1 GitHub source and the AIUC-1 website changelog and produces an automated analysis of how the standard changes between versions, published as static HTML. The audience is AI governance, assurance and compliance practitioners tracking the standard before or during adoption.

Numbers are derived at runtime from the source repository and reconciled against the published website narrative. Snapshots are anchored to git commits because the source publishes no version tags.

For each transition, three views are produced:

• Requirement-level changes. Each requirement (stable ID) is classified as added, removed, identical, or modified. Modifications are further graded by how substantial the wording change is — editorial, clarification, or substantive.
• Control composition. The count of should (normative) and may (informative) control activities per requirement and per principle, with the delta and the should:may ratio. The ratio trajectory shows whether the standard is shifting toward more mandatory or more discretionary controls.
• Per-bullet matching. Individual control activities are matched across releases by text similarity, so that reworded, added, removed, and merged bullets can be distinguished. This view is presented with confidence bands and is treated as indicative only on consolidation releases where rewrites and merges make matching less reliable.

The AIUC-1 source carries no native change stamp, so the substantive/clarification/editorial grading thresholds are currently provisional. Where the website narrative diverges in scope from the requirement-body text, this is flagged on the relevant transition page.

Sources and use

The AIUC-1 source carries no stated licence. This report is transformative — changes, deltas and classifications with bounded snippets only. It is not a substitute for the standard; read the source directly for the canonical text.

Release snapshots

Control composition trajectory

Per-principle change nature over time

Each principle group shows one stacked bar per transition (oldest → newest, left to right). Stack layers count how many requirements under that principle were identical, editorially reworded, clarified, or substantively changed. Added and removed requirements are not stacked here (they appear on per-transition pages).

Transitions

About

AIUC1explorer is an analytical tool and publication by axigetik, a specialist provider of AI security, trust, assurance and risk services. You can contact us via hello@aiuc1explorer.org or sign up to our low-volume announcements list.