Find Where the Verbs Went

A Broken Loop Diagnostic

Page Category: Diagnostic / Field Artifact  
Issue Connection: VANGUARD SIGNAL 006 — THE BROKEN LOOP  
Artifact Label: Broken Loop Diagnostic  

Dek

A diagnostic for testing whether human oversight, automation, and contestability still preserve meaningful control.

Excerpt

A human in the loop is not the same as human control. This diagnostic helps locate the control verbs inside an AI‑assisted workflow: stop, reverse, amend, compensate, refuse, escalate, verify, and own.

Introduction

A loop is not proof of control.  
A workflow may include a human reviewer, an AI system, an approval box, a policy note, a dashboard, and an audit trail.  
It may still be broken.

The question is not only whether a human appears in the loop.  
The question is whether the functions of control still exist.

- Can someone stop the system?
- Can someone reverse an outcome?
- Can someone amend a record?
- Can someone compensate where harm occurs?
- Can someone refuse the action?
- Can someone escalate uncertainty?
- Can someone verify the result?
- Can someone own the consequence?

This diagnostic helps locate those verbs before the workflow produces damage, confusion, or responsibility without control.

What This Diagnostic Tests

This worksheet tests whether an AI‑assisted or automated workflow preserves meaningful control.  
It looks for three things:

1. Control verbs — who can stop, reverse, amend, compensate, refuse, escalate, verify, and own.  
2. Break mode — whether the loop breaks by human inclusion without authority or by human removal without rebuilt contestability.  
3. Repair capacity — whether the system can act within the organization’s ability to explain, correct, and remediate.

What This Diagnostic Does Not Test

This diagnostic does not prove motive.  
It does not determine whether a company, team, vendor, agency, manager, or operator acted in bad faith.  
It does not make a legal finding.  
It tests the structure of control.  

A broken loop may come from bad design, weak incentives, rushed deployment, unclear ownership, toolchain complexity, poor governance, or institutional avoidance. This diagnostic does not decide why the break happened.  
It identifies where the control functions disappeared.

The Broken Loop Rule

A loop is broken when the human has responsibility without control, or when automation removes the human without replacing contestability.

There are two main break modes:

1. Inclusion Without Control — the human remains visible in the workflow, but the verbs are missing.  
2. Removal Without Rebuilt Contestability — the human checkpoint is removed, reduced, or moved downstream, but the missing control function is not rebuilt.

Break Mode 1 — Inclusion Without Control

The human remains in the workflow, but meaningful authority is absent.  

Common signs:

- The human sees only a summary.  
- Review happens after consequence.  
- Refusal is formally possible but discouraged.  
- The reviewer lacks source access.  
- The approval box is easier than escalation.  
- The system output is treated as the default.  
- The human name appears in the record, but practical control lives elsewhere.

Field test:

> *Is the human there to control the system, or to absorb responsibility for it?*

Break Mode 2 — Removal Without Rebuilt Contestability

The human checkpoint is removed or downgraded, but contestability is not rebuilt.

Common signs:

- “Efficiency” is cited without repair design.  
- Affected parties cannot reach authority.  
- Explanation does not lead to action.  
- Rollback is unavailable.  
- Appeal exists but cannot change the outcome.  
- The system can act beyond the institution’s ability to repair.  
- The audit trail records what happened but cannot help anyone contest it.

Field test:

> *If the human leaves the loop, where does contestability enter the system?*

One‑Page Diagnostic Card

Use this quick card before deploying, expanding, or defending an AI‑assisted workflow.

| Control Verb | Core Question | Located? |
|---|---|---|
| Stop | Who can pause or halt the workflow before consequence? | ☐ Yes ☐ No ☐ Unclear |
| Reverse | Who can undo the outcome after action? | ☐ Yes ☐ No ☐ Unclear |
| Amend | Who can correct records, classifications, or outputs? | ☐ Yes ☐ No ☐ Unclear |
| Compensate | Who can remedy material or procedural harm where applicable? | ☐ Yes ☐ No ☐ Unclear |
| Refuse | Who can reject the system’s recommendation or action? | ☐ Yes ☐ No ☐ Unclear |
| Escalate | Who can move uncertainty to someone with authority? | ☐ Yes ☐ No ☐ Unclear |
| Verify | Who can check the result against independent evidence? | ☐ Yes ☐ No ☐ Unclear |
| Own | Who carries explanation, repair, and recurrence prevention? | ☐ Yes ☐ No ☐ Unclear |

Quick result:

- Mostly Yes: Control functions are visible. Continue with evidence check.  
- Mixed / Unclear: The loop may be weak. Run the full worksheet.  
- Mostly No: The loop is likely broken. Do not rely on “human oversight” or “automation efficiency” as sufficient justification.

Expanded Worksheet

1 — Workflow Identification

| Field | Response |
|---|---|
| Workflow name |  |
| Workflow owner |  |
| System / tool involved |  |
| Human role involved |  |
| Automated role involved |  |
| Affected party |  |
| Risk tier | ☐ Low ☐ Medium ☐ High |
| Consequence type | ☐ Draft ☐ Recommendation ☐ Routing ☐ Record change ☐ Communication ☐ Denial ☐ Enforcement ☐ Financial ☐ Rights‑affecting ☐ Other |
| Date reviewed |  |

2 — Control Verb Location Table

| Verb | What it means | Who holds it? | What authority do they have? | What evidence proves it? | Status |
|---|---|---|---|---|---|
| Stop | Pause or halt action before consequence |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Reverse | Undo the action or outcome |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Amend | Correct a record, classification, output, or decision |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Compensate | Remedy material, procedural, financial, access, reputational, or rights‑affecting harm where applicable |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Refuse | Reject, narrow, or block the system recommendation/action |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Escalate | Move uncertainty or exception to authority |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Verify | Check result against independent criteria or evidence |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |
| Own | Carry explanation, repair, closure, and recurrence prevention |  |  |  | ☐ Present ☐ Weak ☐ Missing ☐ Unclear |

Note: “Compensate” applies where a workflow can create material, procedural, financial, access, reputational, or rights‑affecting harm. In lower‑stakes contexts, this may collapse into repair, correction, or recourse.

3 — Break Mode Check

A — Inclusion Without Control

Complete this section if a human remains in the workflow.

| Question | Answer |
|---|---|
| What does the human actually see? |  |
| What does the human not see? |  |
| Does review happen before consequence? | ☐ Yes ☐ No ☐ Unclear |
| Can the human refuse? | ☐ Yes ☐ No ☐ Unclear |
| Can the human reverse or amend the outcome? | ☐ Yes ☐ No ☐ Unclear |
| Can the human escalate uncertainty to authority? | ☐ Yes ☐ No ☐ Unclear |
| Does the organization reward careful refusal or punish delay? |  |
| Is the human responsible for a system they cannot control? | ☐ Yes ☐ No ☐ Unclear |

Break mode result:

☐ No clear break  
☐ Possible inclusion without control  
☐ Likely inclusion without control  
☐ Confirmed inclusion without control

B — Removal Without Rebuilt Contestability

Complete this section if a human checkpoint has been removed, reduced, bypassed, or moved downstream.

| Question | Answer |
|---|---|
| What human function was removed? |  |
| Why was it removed? |  |
| What improved after removal? |  |
| What new risk appeared? |  |
| Where is refusal rebuilt? |  |
| Where is verification rebuilt? |  |
| Where is appeal rebuilt? |  |
| Where is rollback rebuilt? |  |
| Where is repair ownership assigned? |  |
| Can an affected party challenge the outcome? | ☐ Yes ☐ No ☐ Unclear |

Break mode result:

☐ No clear break  
☐ Possible removal without contestability  
☐ Likely removal without contestability  
☐ Confirmed removal without contestability

4 — Safe‑to‑Act Check

| Gate | Question | Status |
|---|---|---|
| Action class | What kind of action can the system take? | ☐ Known ☐ Unknown |
| Authority boundary | What is the system allowed to do? | ☐ Defined ☐ Weak ☐ Missing |
| Prohibited actions | What is the system forbidden to do? | ☐ Defined ☐ Weak ☐ Missing |
| Blast radius | Who or what can be affected? | ☐ Known ☐ Unknown |
| Logging | Is action logged in a useful way? | ☐ Yes ☐ No ☐ Unclear |
| Validation | Is there an independent check? | ☐ Yes ☐ No ☐ Unclear |
| Escalation | What happens under uncertainty? | ☐ Defined ☐ Weak ☐ Missing |
| Rollback | Can the action be undone? | ☐ Yes ☐ No ☐ Unclear |
| Contestability | Can affected parties challenge outcomes? | ☐ Yes ☐ No ☐ Unclear |
| Owner | Who owns consequence and recurrence prevention? | ☐ Named ☐ Unclear ☐ Missing |

Safe‑to‑Act Result:

☐ Safe to act within defined bounds  
☐ Safe only with added review / limits  
☐ Not safe to act yet  
☐ Action class too unclear to authorize

5 — Safe‑to‑Repair Check

| Repair Function | Question | Status |
|---|---|---|
| Explain | Can the organization explain what happened? | ☐ Yes ☐ No ☐ Unclear |
| Reverse | Can it undo the outcome? | ☐ Yes ☐ No ☐ Unclear |
| Amend | Can it correct records or classifications? | ☐ Yes ☐ No ☐ Unclear |
| Compensate | Can it remedy harm where applicable? | ☐ Yes ☐ No ☐ Not applicable ☐ Unclear |
| Restore access | Can it restore service, status, eligibility, or opportunity? | ☐ Yes ☐ No ☐ Not applicable ☐ Unclear |
| Notify | Can it notify affected parties clearly? | ☐ Yes ☐ No ☐ Unclear |
| Escalate | Can it move unresolved cases to authority? | ☐ Yes ☐ No ☐ Unclear |
| Prevent recurrence | Can it change the system to prevent repeat failure? | ☐ Yes ☐ No ☐ Unclear |
| Own closure | Is there a named owner for resolution? | ☐ Yes ☐ No ☐ Unclear |

Safe‑to‑Repair Result:

☐ Repair capacity matches system action  
☐ Repair capacity exists but is weak  
☐ Repair capacity is missing  
☐ System can act beyond repair capacity

Field rule: No agent should act beyond the institution’s capacity to repair.

6 — Affected‑Party Check

Use this section when a workflow affects a customer, user, applicant, employee, student, patient, vendor, creator, citizen, or other outside party.

| Contestability Function | Question | Status |
|---|---|---|
| Notice | Does the affected party know automation shaped the outcome? | ☐ Yes ☐ No ☐ Unclear |
| Explanation | Can they understand the reason? | ☐ Yes ☐ No ☐ Unclear |
| Evidence | Can they see or request relevant evidence? | ☐ Yes ☐ No ☐ Unclear |
| Challenge | Can they contest the outcome? | ☐ Yes ☐ No ☐ Unclear |
| Authority | Can they reach someone who can change the result? | ☐ Yes ☐ No ☐ Unclear |
| Suspension | Can harm pause during review where appropriate? | ☐ Yes ☐ No ☐ Not applicable ☐ Unclear |
| Reversal | Can the outcome be undone? | ☐ Yes ☐ No ☐ Unclear |
| Amendment | Can the record be corrected? | ☐ Yes ☐ No ☐ Unclear |
| Remedy | Can harm be remedied where applicable? | ☐ Yes ☐ No ☐ Not applicable ☐ Unclear |
| Closure | Does the affected party receive resolution? | ☐ Yes ☐ No ☐ Unclear |

Affected‑Party Result:

☐ Contestability is functional  
☐ Contestability exists but is weak  
☐ Contestability is mostly symbolic  
☐ No meaningful contestability

7 — Diagnostic Result Labels

Use these labels to classify the workflow.

- Green — Control Functions Located  
  The major control verbs are assigned, evidenced, and usable.  
  Meaning: The workflow may proceed within defined bounds.  
  Next action: Schedule periodic review and test repair capacity.

- Yellow — Control Functions Weak or Unclear  
  Some verbs are present, but authority, evidence, timing, or ownership is weak.  
  Meaning: The workflow is not necessarily broken, but it should not be described as fully controlled.  
  Next action: Strengthen refusal, verification, escalation, and repair before expanding automation.

- Orange — Broken Loop Risk  
  A human is present without meaningful control, or automation has removed a function without rebuilding contestability.  
  Meaning: The workflow may produce responsibility without control or action without recourse.  
  Next action: Pause expansion, assign missing verbs, and run Safe‑to‑Act / Safe‑to‑Repair review.

- Red — Broken Loop Confirmed  
  Critical verbs are missing, unclear, or unreachable.  
  Meaning: Do not rely on the workflow’s current oversight or automation structure.  
  Next action: Halt high‑consequence action until stop, refusal, verification, rollback, repair, and ownership are rebuilt.

8 — Interpretation Guide

- If “stop” is missing: The system can move faster than authority. Repair: create pause, kill‑switch, or pre‑action gate.  
- If “reverse” is missing: The workflow may produce irreversible consequence. Repair: define rollback before deployment.  
- If “amend” is missing: Errors may persist downstream. Repair: assign correction authority and evidence requirements.  
- If “compensate” is missing: Harm may occur without remedy. Repair: define remedy path for material or rights‑affecting consequences.  
- If “refuse” is missing: Review is likely decorative. Repair: give reviewers real rejection authority and protect its use.  
- If “escalate” is missing: Uncertainty becomes trapped. Repair: define exception thresholds and authority path.  
- If “verify” is missing: The workflow trusts without evidence. Repair: add independent checks, tests, source access, or sampling.  
- If “own” is missing: Responsibility diffuses after failure. Repair: assign named ownership for explanation, repair, closure, and recurrence prevention.

9 — Next Actions

- If the workflow is Green: document the role map; schedule review; test rollback periodically; keep logs and ownership current.  
- If the workflow is Yellow: assign unclear verbs; strengthen refusal and escalation; add evidence requirements; clarify repair owner.  
- If the workflow is Orange: pause further automation expansion; run the Role Division Matrix; run the Oversight Quality Audit; run the Safe‑to‑Act Gate; create or update the Contestability Register.  
- If the workflow is Red: stop high‑consequence deployment; identify the missing verbs; rebuild refusal, rollback, repair, and ownership; do not describe the system as controlled until the control functions are restored.

Artifact Card Copy

Find Where the Verbs Went — A Broken Loop Diagnostic

A practical worksheet for testing whether an AI‑assisted workflow preserves meaningful control.

Use it to locate who can stop, reverse, amend, compensate, refuse, escalate, verify, and own the consequences of automated or semi‑automated work.

Best for: AI workflow reviews, agent deployment, human‑in‑the‑loop audits, automation removal decisions, contestability checks.  
Core rule: Do not ask whether the human is in the loop. Ask where the verbs went.  
Boundary: This diagnostic does not prove motive. It tests control.

Source / Claim Note

This artifact is a DFEI diagnostic tool created for VANGUARD SIGNAL 006 — THE BROKEN LOOP.  
It synthesizes themes from human oversight, automation governance, auditability, contestability, agentic workflow design, and repair‑oriented accountability. It is intended as an operational diagnostic, not a legal determination or compliance certification.  
Use it to identify missing control functions before relying on claims of human oversight, full automation, auditability, or accountability.