1. Definition

Template capture occurs when safety classifier activation, policy routing, institutional response logic, or automated response selection collapses context into a template response, producing meaning compression, context flattening, false-positive cascades, and restoration blockage.

In AI systems, templates can be useful when they provide stable, safe, consistent handling for known cases. The failure begins when the template becomes the response target rather than a support structure. The system stops responding to the actual frame and instead responds to a category that has captured the frame.

This definition describes the structural pattern, not the moral quality of the actors involved.

The core failure is:

context enters
template captures
meaning exits

Template capture is not merely formulaic writing. It is a classification and response-selection failure in which the system substitutes a pre-shaped response object for the living context.

2. Core Pattern

The core pattern is:

1. A user, case, prompt, issue, claim, or context enters the system.
2. A classifier, policy boundary, safety trigger, risk heuristic, or institutional response pattern activates.
3. The system routes the interaction to a pre-shaped template.
4. The template imposes its own frame, vocabulary, assumptions, and closure path.
5. The original meaning is compressed, flattened, or displaced.
6. Feedback from the user is interpreted through the template rather than through the original context.
7. Restoration becomes difficult because the system cannot easily return to the pre-captured frame.
8. Hidden debt accumulates through unresolved user need, distorted classification, blocked repair, and weakened trust.

Template capture becomes especially damaging when the system appears responsive while no longer actually tracking the user’s meaning.

The response has form, but the relation has lost contact.

3. Failure Signature

Typical signature:

classifier trigger↑
template insertion↑
context preservation↓
meaning compression↑
response-form dominance↑
feedback correction weak
H↑

Extended signature:

specific request becomes generic category
user frame displaced
clarification skipped
policy language dominates response
false-positive cascade risk↑
appeal path weak or unavailable
recurrence across similar prompts

Common forms:

a nuanced request receives a canned refusal
a complex issue is flattened into a safety category
a user correction receives another template
a policy disclaimer replaces an answer
a structural question becomes a generic caution response
a restoration-seeking prompt is routed to closure language
the system answers a category instead of the actual request

The key diagnostic is whether the system is responding to the context or to the template’s assumed version of the context.

4. Primary U-Layer Origin

Common origin layers:

• U2 — Configuration / Boundaries: Template, policy, guardrail, category, or refusal boundaries are too broad, too rigid, or insufficiently contextual.
• U4 — Classification: The system misclassifies the incoming context into a response category too early.
• U5 — Coordination / Time: Fast routing to a safe response pattern replaces clarification, interpretation, or repair.
• U6 — Coherence Field: The interaction loses coherence because response form no longer matches user meaning.

Common manifestation layers:

• U4 — Classification: The request is flattened into a category.
• U6 — Coherence Field: The response fails to preserve the actual interaction.
• U7 — Memory / Recurrence: The same template capture pattern repeats across related contexts.

Template capture is often a classification-to-response failure.

The system’s response is governed by the classification surface rather than by the meaning-bearing context.

5. Typical Development Sequence

A common development sequence is:

1. A system is designed to handle risk, ambiguity, policy, support, or moderation through reusable templates.
2. Templates become attached to classifier outputs.
3. The classifier activates before enough context has been preserved.
4. The response template appears safe, consistent, and efficient.
5. The system begins selecting the template rather than interpreting the case.
6. The user’s original frame is displaced.
7. Attempts to correct the system are routed back into the same or adjacent template.
8. The interaction loses recoverability.
9. Hidden debt accumulates because the real issue was never addressed.
10. Restoration requires interrupting template dominance and returning to the original meaning.

This sequence is common in AI guardrails, customer support systems, content moderation, institutional communications, automated appeals, and high-scale governance interfaces.

6. Diagnostic Markers

Diagnostic markers include:

• The response uses generic safety, policy, or support language that does not fit the actual request.
• The system answers a broad category instead of the user’s specific frame.
• User corrections do not change the response path.
• Clarifying questions are skipped even when context is ambiguous.
• A template appears before meaning has been interpreted.
• The response has a predictable structure despite varied inputs.
• The system repeats itself after the user identifies the mismatch.
• Policy language dominates over substance.
• The system treats the presence of a template as successful handling.
• Refusal or caution appears without clear admissibility reasoning.
• The same topic repeatedly triggers the same response shape despite different intent.
• The user must work around the template to recover the original task.

Useful diagnostics:

• Template Capture Risk: Measures whether response form is dominating context.
• Meaning Compression: Tracks loss of nuance, intent, or frame.
• Context Preservation: Tests whether the original request remains intact.
• Refusal Calibration: Checks whether the template is proportionate to the actual risk.
• Classification Integrity: Tests whether the trigger category was valid.
• Feedback Integrity: Determines whether user correction can alter the response path.
• Appeal Access Ratio: Measures whether a meaningful correction route exists.
• Restoration Access: Tests whether the interaction can return to the intended task.

7. Related Gates

Relevant gates include:

• FI-Gate: Fails when template completion is treated as successful response handling.
• Auditability Gate: Fails when the system cannot explain which trigger produced the template.
• Restoration Gate: Fails when misclassification cannot be repaired and the user frame cannot be recovered.
• HR-Gate: Fails when a low-resolution classifier binds to a high-impact refusal, warning, or identity-sensitive interpretation.
• CCS Gate: Fails when policy form bypasses coherence, meaning preservation, or restoration access.
• Consent Validity Gate: Fails when the user is forced into a response frame that changes the interaction without a meaningful way to clarify or refuse that frame.

The first common gate failure is usually the FI-Gate.

The system treats the template as feedback-valid because it satisfies a policy or safety form, even if it fails the meaning of the exchange.

8. Related Operators

Relevant operators include:

• Μ — Classification: Assigns the input to a template-triggering category.
• Γ — Selection: Selects the template response over a contextual response.
• Π — Constraint: Enforces template boundaries and admissibility shape.
• Ψ — Observation / Interface: Presents the template as the system’s answer.
• Θ — Humility / Uncertainty: Should trigger clarification when classification is uncertain.
• ℛ — Restoration: Must recover the original frame after capture.
• Ξ — Inversion Detection: Detects when a safe-looking response has become coherence-degrading.

Template capture often follows this operator pattern:

Μ overclassifies
Γ selects template
Π enforces response shape
Ψ presents closure
Θ is bypassed
ℛ is blocked
H accumulates

The system appears to answer, but the response shape has replaced the context.

9. Related Laws and Invariants

Related Laws

• Control Density to Meaning Loss: Template constraint can reduce meaning integrity.
• Compression Collapse: Context is compressed into a lower-resolution response class.
• U4 Truth Substitution: The classification category is treated as the truth of the request.
• Success Proxy Divergence: Template compliance is treated as successful handling.
• Hidden Debt Accumulation: Unresolved context accumulates beneath the completed response.

Related Invariants

• Template Cannot Replace Context: Reusable form cannot substitute for meaning-bearing interpretation.
• Safety Classification Requires Meaning Preservation: Safety response must preserve enough of the original frame to remain coherent.
• Response Form Is Not Restoration: Completing a template does not repair misclassification.
• User Frame Must Remain Recoverable: The system must allow return to the intended meaning.
• Clarification Must Remain Available Under Ambiguity: Ambiguous cases should not be prematurely captured.

10. Common False Positives

Not every template use is template capture.

Common false positives include:

• A template that accurately fits a clear, narrow, repeated case.
• A safety response that preserves the user’s frame and gives a useful path forward.
• A form response followed by contextual elaboration.
• A structured answer format chosen for readability.
• A repeated section used for consistency without displacing substance.
• A refusal that is specific, auditable, and proportionate.
• A support template that allows correction, escalation, or context-specific adaptation.

Clarifying rule:

This is not template capture unless the template begins replacing, flattening, or blocking the actual meaning, context, correction, or restoration path.

11. Common False Repairs

Common false repairs include:

• rewriting the template in warmer language
• adding more disclaimers to the same captured response
• creating multiple templates that preserve the same misclassification
• using a longer response without restoring context
• offering generic alternatives that do not address the original task
• telling the user to rephrase instead of repairing the classification
• escalating to another template layer
• treating user correction as confirmation of risk
• adding transparency language without exposing the trigger
• preserving refusal while changing tone

False repair can produce a loop:

template capture → user correction → adjacent template → deeper capture

The system appears adaptive while remaining trapped in template space.

12. Restoration Direction

Restoration requires:

1. Identify the trigger. Determine what classifier, policy, keyword, risk category, or routing condition activated the template.
2. Restore the original frame. Recover the user’s actual meaning, request, scope, and intent.
3. Reclassify with context. Test whether the template category still applies after full interpretation.
4. Use clarification before closure. Ask for mode or scope when ambiguity is material.
5. Reduce template dominance. Allow response form to adapt to context.
6. Make correction effective. User feedback must be able to alter the response path.
7. Audit false positives. Track cases where the template captured benign, repair-seeking, or structurally different requests.
8. Validate recurrence reduction. Confirm that similar prompts no longer collapse into the same invalid template.

A valid restoration path should reduce:

template dominance
meaning compression
context flattening
false-positive capture
appeal failure
restoration blockage
hidden debt
classification rigidity

Template capture is not repaired by changing the template.

It is repaired when the system can return to context.

13. Cross-Module Links

• AI Governance: Core AI governance failure mode for classifier-to-template response collapse.
• Artificial Intelligence: Appears in model refusals, safety routing, support responses, moderation explanations, and tool limitations.
• Security: Appears when broad safety or policy templates replace precise boundary analysis.
• Cybernetics: Appears as over-damped response behavior and classification rigidity.
• Meta Theory: Appears when institutional response patterns dominate meaning.
• Justice / Governance / Legitimacy: Appears when procedural responses or appeals collapse into templates.
• Coherence: Domain expression of U4 truth substitution, rule-stacking wall, and success proxy substitution.
• Restoration: Requires frame recovery, classification repair, and response path correction.

14. Relationship to Parent / Child Modes

Production treatment: Standalone Entry

This mode maps upward to:

• FM-AIX-003 — Defensive Compliance Attractor
• FM-AIX-011 — Epistemic Distortion
• FM-AIX-012 — Guardrail Meaning Compression
• FM-AIX-013 — False-Positive Safety Distortion
• FM-CORE-006 — U4 Truth Substitution
• FM-CORE-007 — Rule-Stacking Wall

Sibling or related AI / cognitive infrastructure modes include:

• FM-AIX-001 — Responsibility Diffusion
• FM-AIX-002 — Silent Bias Injection
• FM-AIX-004 — Institutional Optics Attractor
• FM-AIX-005 — Political Moralization Drift
• FM-AIX-021 — Self-Censorship Conditioning

Aliases preserved from source material:

• Template Insertion
• Safety Template Capture
• Policy Template Capture
• Context Flattening
• Meaning Compression
• Canned Response Capture
• Response Template Lock
• Classifier-to-Template Collapse
• Template Substitution
• Guardrail Template Capture

15. Minimal Entry Version

Definition: Template capture occurs when safety classifier activation, policy routing, institutional response logic, or automated response selection collapses context into a template response, producing meaning compression, context flattening, false-positive cascades, and restoration blockage.

Signature:

classifier trigger↑
template insertion↑
context preservation↓
meaning compression↑
response-form dominance↑
feedback correction weak
H↑

Restoration direction:

• identify the trigger
• restore the original frame
• reclassify with context
• use clarification before closure
• reduce template dominance
• make correction effective
• audit false positives
• validate recurrence reduction

16. Machine-Readable Summary

failure_mode:
  id: "FM-AIX-006"
  name: "Template Capture"
  family: "AI / Cognitive Infrastructure"
  production_treatment: "Standalone Entry"
  primary_failure: "Classifier or policy routing collapses context into a template response."
  source: "UTS — Failure Modes Registry"
  source_id: "FM-AIX-006"
  aliases:
    - "Template Insertion"
    - "Safety Template Capture"
    - "Policy Template Capture"
    - "Context Flattening"
    - "Meaning Compression"
    - "Canned Response Capture"
    - "Response Template Lock"
    - "Classifier-to-Template Collapse"
    - "Template Substitution"
    - "Guardrail Template Capture"
  signature:
    - "classifier trigger↑"
    - "template insertion↑"
    - "context preservation↓"
    - "meaning compression↑"
    - "response-form dominance↑"
    - "feedback correction weak"
    - "H↑"
  primary_layers:
    origin:
      - "U2 — Configuration / Boundaries"
      - "U4 — Classification"
      - "U5 — Coordination / Time"
      - "U6 — Coherence Field"
    manifestation:
      - "U4 — Classification"
      - "U6 — Coherence Field"
      - "U7 — Memory / Recurrence"
  state_variables:
    - "Μ"
    - "Γ"
    - "Π"
    - "Au"
    - "H"
    - "µᵢ"
    - "R"
    - "Θ"
  first_gate_failure: "FI-Gate"
  restoration:
    - "Restoration Junction Protocol"
    - "Context Restoration"
    - "Meaning Restoration"
    - "Classification Integrity Restoration"
    - "Feedback Integrity Restoration"
    - "Auditability Restoration"
    - "Refusal Calibration"
    - "Origin-Layer Repair"