---
schema_version: "1.0"
id: "GL-252"
title: "GL-252 — Signed Decision Provenance"
slug: "gl-252-signed-decision-provenance"
type: "glossary_term"
status: "draft"
version: "0.1.0"
last_updated: "2026-06-24"
summary: "Signed Decision Provenance is a traceable record of who or what made, approved, changed, constrained, or automated a decision, with cryptographic or governance-backed integrity."
canonical_url: "/archive/glossary/registry/gl-252-signed-decision-provenance"
citation_id: "gl-252-signed-decision-provenance-v0-1-0"
canon:
tier: "registry"
state: "draft"
source: "UTS Glossary Simplified Registry"
source_id: "GL-252"
classification:
family: "Glossary"
module: "AI Governance Terms"
module_group: "Reference Systems"
density: "Reference"
audience:
- "UTS readers"
- "researchers"
- "builders"
- "AI readers"
- "machine readers"
tags:
- "glossary"
- "registry"
- "gl-252"
- "signed-decision-provenance"
- "provenance"
- "auditability"
aliases:
- "Signed Decision Provenance"
- "Decision provenance"
- "Signed provenance"
- "Traceable decision record"
related:
laws:
- "Temporal Audit Asymmetry"
- "Guardrails as Epistemic Infrastructure"
- "Hidden Debt Return Law"
invariants:
- "O ≠ Φ"
operators:
- "Au"
- "Μ"
- "Π"
- "Τ"
- "ℛ"
gates:
- "Au-Actuation"
- "FI-Gate"
- "HR-Gate"
- "R Sufficiency"
- "Τ Validation"
diagnostics:
- "Au"
- "FI"
- "H"
- "O"
- "provenance_integrity"
- "legitimacy"
- "recurrence"
failure_modes:
- "Provenance Collapse"
- "Interface Capture"
- "Auditability Collapse"
- "Paper Governance"
- "Evaluator Capture"
restoration_arcs:
- "Legibility Restoration"
- "Truth Reconstruction"
- "Repair First AI Architecture"
- "Temporal Proof Arc"
modules:
- "glossary"
- "ai-governance"
terms:
- "Authority Registry"
- "Tamper-Evident Audit Trail"
- "Cognitive Infrastructure Governance"
- "Auditability"
- "Causal Traceability"
navigation:
order: 252
parent: "glossary"
visible: true
provenance:
created_from: "glossary-simplified-continuation"
source_thread: "GLOSSARY-REFACTOR.md"
source_file: "glossary-raw.docx"
notes: "Continued AI Governance Terms sequence."
entry:
term_id: "GL-252"
term: "Signed Decision Provenance"
term_class:
- "AI Governance Term"
- "Provenance Mechanism"
- "Decision Traceability Structure"
symbols:
- "Au"
- "Τ"
---1. Short Definition
Signed Decision Provenance is a traceable record of who or what made, approved, changed, constrained, or automated a decision, with cryptographic or governance-backed integrity.
2. Canonical Definition
In UTS, Signed Decision Provenance ensures that consequential decisions cannot detach from their origin, authority, rationale, constraint set, and change history.
Canonical purpose:
decision made
→ origin signed
→ authority visible
→ rationale traceable
→ repair assignableFor AI governance, signed provenance applies to:
- model releases
- guardrail changes
- moderation decisions
- dataset changes
- policy changes
- ranking changes
- capability gate approvals
- automated classifications
- escalations
- overrides
- user-impacting restrictions
Signed Decision Provenance prevents authority, causality, and repair obligation from dissolving into system complexity.
3. Functional Role in UTS
Signed Decision Provenance supports:
- auditability
- AI governance
- public trust
- incident response
- appeal systems
- accountability
- model change control
- guardrail review
- data governance
- legitimacy restoration
- recurrence prevention
It makes consequential decisions inspectable after the fact.
4. Diagnostic Signatures
Provenance active
decision origin visible
authority linked
rationale recorded
constraints recorded
signature valid
Au↑
repair assignableProvenance collapse
decision affects users
but origin, authority, or rationale cannot be traced
H↑
legitimacy↓Provenance restored
decision chain reconstructed
authority registry linked
audit trail intact
appeal path usable5. Canonical Distinctions
Signed Decision Provenance is not logging alone
Logs may record events without authority, rationale, or integrity.
Signed Decision Provenance is not transparency theater
It must support actual audit and repair.
Signed Decision Provenance is not public exposure of everything
Valid privacy boundaries can remain while decision authority is traceable.
Signed Decision Provenance is not blame
It enables responsibility mapping before consequence allocation.
6. U-Layer Mapping
| U-Layer | Signed Decision Provenance Expression |
|---|---|
| U0 | Infrastructure and deployment changes are signed. |
| U1 | Resource allocation and compute/data access decisions are traceable. |
| U2 | Permission, boundary, and authority decisions are signed. |
| U3 | Runtime decisions and overrides are logged with provenance. |
| U4 | Classification, moderation, and policy decisions are traceable. |
| U5 | Decision timing and change sequence are preserved. |
| U6 | Field-impacting decisions remain auditable. |
| U7 | Historical decision memory is preserved. |
| U8 | External authority and regulatory decisions are linked. |
7. Common Failure Patterns Addressed
| Failure Pattern | Description |
|---|---|
| Provenance Collapse | Decisions cannot be traced to origin or authority. |
| Interface Capture | Interface mediates decisions without visible authority. |
| Auditability Collapse | Causal path cannot be reconstructed. |
| Paper Governance | Governance claims exist without decision traceability. |
| Responsibility Diffusion | Repair burden cannot be assigned. |
8. Restoration Implications
Signed Decision Provenance must be built into high-impact decision surfaces.
Typical sequence:
Μ identify consequential decision surfaces
→ link each decision to authority registry
→ record rationale and constraint set
→ sign decision artifact
→ preserve tamper-evident audit trail
→ expose review pathway
→ assign repair obligation where needed
→ Τ validate provenance integrityProvenance is coherent when decisions remain traceable enough to audit, challenge, correct, and repair.
9. Machine-Readable Summary
glossary_entry:
id: "GL-252"
term: "Signed Decision Provenance"
symbols:
- "Au"
- "Τ"
short_definition: "A traceable record of who or what made, approved, changed, constrained, or automated a decision, with cryptographic or governance-backed integrity."
term_family: "AI Governance Terms"
term_class:
- "AI Governance Term"
- "Provenance Mechanism"
- "Decision Traceability Structure"
canonical_purpose:
- "decision made → origin signed → authority visible → rationale traceable → repair assignable"
diagnostic_positive:
- "decision origin visible"
- "authority linked"
- "rationale recorded"
- "constraints recorded"
- "signature valid"
- "Au↑"
- "repair assignable"
diagnostic_negative:
- "decision affects users"
- "origin cannot be traced"
- "authority cannot be traced"
- "rationale cannot be traced"
- "H↑"
- "legitimacy↓"
restoration_requirements:
- "consequential decision surface identification"
- "authority registry linking"
- "rationale and constraint recording"
- "decision artifact signing"
- "tamper-evident audit trail"
- "review pathway exposure"
- "repair obligation assignment"
- "time validation"