Workbench / Issues

Policy Update Not Reflected in Output

A policy, rule, standard, or instruction has been updated, but the AI output still follows the older version.

What This Looks Like

A policy, rule, standard, guideline, rubric, or instruction has been updated, but the AI output still reflects the older version. The user may see outdated language, old thresholds, prior approval rules, obsolete categories, or earlier workflow requirements appear after the update should have taken effect.

Why It Matters

Policy updates are only useful if they change the outputs they are supposed to govern. When AI output continues to follow an older policy, users may believe the current policy is being applied when it is not. This creates audit, compliance, and operational risk, especially when the old and new policies lead to different decisions.

Structural Signal

The declared governing rule has changed, but the generated output is still governed by a prior rule state. The issue is not just that the answer is outdated; it is that policy version and output behavior are no longer aligned.

Common Triggers

When to Use This Issue

Use this Issue when a policy or rule has changed and the AI output still follows the older policy state.

When Not to Use This Issue

Do not use this Issue when the policy is merely unclear or when the AI violates a policy that has not changed. Use this Issue when the update itself is central: the output fails to reflect the new governing rule.

Category

Rules & Policies

Primary Pattern

PAT-0190 — Contract Drift

Declared Patterns

Derived Primary Lenses

Derived Secondary Lenses

Search Intents

Ontology Metadata

Code
ISS-0042
Version
ISS-0042@0.1.0
Ontology release
0.1.0
Updated
2026-05-10T00:00:00Z

History

No public history entries recorded.

View full ontology changelog