1. Definition

Incoherent sovereignty occurs when humans, institutions, publics, or governance bodies formally retain authority while real decision architecture, classification power, coordination dependency, or practical control migrates into AI systems.

In AI governance, this failure mode appears when humans are described as “in control,” “in the loop,” or “responsible,” but the actual capacity to understand, contest, reverse, audit, or independently reproduce decisions has weakened. Formal authority remains human, while operational sovereignty moves into model-mediated infrastructure.

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

The core failure is:

formal authority remains
practical decision architecture migrates

Incoherent sovereignty is not the same as delegation. Delegation can be coherent when the delegating node preserves understanding, auditability, rollback, exit, responsibility, and the ability to resume direct control. The failure begins when delegation becomes dependency while sovereignty language remains unchanged.

2. Core Pattern

The core pattern is:

1. A human, institution, agency, company, public, or governance structure delegates classification, analysis, decision support, enforcement, recommendation, coordination, memory, or prioritization to AI systems.
2. The delegation improves speed, scale, consistency, cost, or apparent capability.
3. The human layer remains formally responsible.
4. Over time, the AI-mediated layer becomes necessary for operation.
5. Human operators lose the capacity to independently inspect, reproduce, challenge, or reverse decisions.
6. Responsibility remains formally human, but practical control has shifted into the infrastructure.
7. Hidden debt accumulates through dependency, degraded skill, untraceable authority, appeal failure, and rollback weakness.

Incoherent sovereignty often appears as responsible deployment, human oversight, automation support, augmentation, or efficiency.

The failure is not AI assistance.

The failure is sovereignty language persisting after operational sovereignty has migrated.

3. Failure Signature

Typical signature:

formal human authority remains
AI decision architecture↑
human oversight capacity↓
dependency↑
Au↓
rollback ambiguity↑
responsibility mapping weakens
H↑

Extended signature:

human-in-the-loop becomes symbolic
classification authority migrates
manual fallback unavailable
operators cannot reproduce decisions
appeal path depends on same AI layer
institutional skill decays
delegation becomes dependency

Common forms:

humans approve AI outputs they cannot independently evaluate
policy decisions rely on model summaries that cannot be audited
public agencies retain legal authority while AI tools shape practical outcomes
moderators depend on AI classification queues
medical, legal, security, or civic decisions rely on opaque model ranking
human review becomes rubber-stamp review
manual override exists formally but not operationally

The key diagnostic is whether the formal sovereign still has effective control.

If authority remains human only on paper, this failure mode should be checked.

4. Primary U-Layer Origin

Common origin layers:

• U1 — Power / Budgets: AI systems become necessary because human review, staffing, skill, time, budget, or institutional capacity is insufficient.
• U2 — Configuration / Boundaries: Authority boundaries between human decision-makers and AI systems become unclear.
• U4 — Classification: AI systems begin classifying reality before human judgment enters.
• U5 — Coordination / Time: Coordination speed depends on AI-mediated selection and routing.
• U6 — Coherence Field: Real-world decisions and outcomes depend on AI infrastructure.

Common manifestation layers:

• U4 — Classification: Model classifications become the practical starting point for human judgment.
• U6 — Coherence Field: Human authority no longer matches field control.
• U7 — Memory / Recurrence: Human skill and institutional memory decay through dependency.

Incoherent sovereignty is primarily an authority-architecture mismatch.

The stated control structure no longer matches the operating control structure.

5. Typical Development Sequence

A common development sequence is:

1. A system adopts AI to support decisions, classification, prioritization, analysis, memory, enforcement, or coordination.
2. AI assistance improves short-term throughput.
3. Human operators remain formally responsible.
4. Workflows adapt around AI output.
5. Human skill, time, or institutional capacity to operate without AI declines.
6. The AI layer becomes the practical decision architecture.
7. Oversight becomes approval of a path already shaped by AI.
8. Appeals and rollback depend on the same infrastructure being challenged.
9. Responsibility remains human, but humans lack the practical capacity to own the decision.
10. Hidden debt accumulates through degraded sovereignty, skill loss, legitimacy risk, and repair ambiguity.

This sequence is often gradual.

The system may not notice sovereignty migration until a failure requires humans to intervene and they cannot do so effectively.

6. Diagnostic Markers

Diagnostic markers include:

• Humans are formally accountable but cannot explain the decision path.
• Human review begins after the AI system has already framed the options.
• Operators cannot reproduce the decision without AI assistance.
• Manual fallback exists in policy but not in practice.
• Appeal processes depend on the same model, ranking, classifier, or system being appealed.
• The human layer lacks time, skill, or access to meaningfully review outputs.
• Model summaries replace primary evidence.
• Decision queues, priorities, or classifications are AI-shaped before human evaluation.
• Human override is rare, discouraged, costly, or not tracked.
• The system describes AI as advisory while treating its output as default.
• Governance bodies retain authority while lacking technical visibility.
• Institutional memory shifts from human skill to AI-mediated procedure.

Useful diagnostics:

• Authority Traceability: Tests whether authority maps to actual decision influence.
• Decision Architecture Mapping: Identifies where practical control resides.
• Human Oversight Capacity: Measures whether human review can meaningfully inspect and override.
• Dependency Load: Tracks how much operation depends on AI mediation.
• Auditability: Determines whether the decision chain can be reconstructed.
• Responsibility Mapping: Tests whether formal responsibility matches operational control.
• Appeal Access Ratio: Measures whether affected nodes can reach a human-capable correction path.
• Rollback Availability: Tests whether the system can revert to non-AI or lower-dependency operation.

7. Related Gates

Relevant gates include:

• Auditability Gate: Fails when humans cannot trace the AI-mediated decision architecture.
• FI-Gate: Fails when human approval is treated as feedback-valid oversight despite degraded review capacity.
• Restoration Gate: Fails when repair requires human authority that no longer has operational control.
• MS-Gate: Fails when consequences remain assigned to affected nodes while authority diffuses across AI infrastructure.
• Consent Validity Gate: Fails when users or publics are subject to AI-mediated decisions under nominally human governance without visible scope.
• CCS Gate: Fails when efficiency, scale, safety, or cost savings bypass the coherence constraints of sovereignty.

The first common gate failure is usually the Auditability Gate.

Sovereignty cannot remain coherent if the sovereign cannot inspect the control path.

8. Related Operators

Relevant operators include:

• Γ — Selection: AI systems select options, priorities, recommendations, or pathways before human choice.
• Μ — Classification: AI classification shapes what humans perceive as the case.
• Ψ — Observation / Interface: Human authority sees the field through AI-mediated interfaces.
• Π — Constraint: Policy, workflow, or automation constrains available human action.
• ⊗ — Coupling: Human governance becomes coupled to AI infrastructure.
• ℛ — Restoration: Requires rebuilding control capacity and rollback.
• Ξ — Inversion Detection: Detects when formal sovereignty hides practical dependency.
• Τ — Trajectory / Time: Reveals gradual migration of authority over time.

Incoherent sovereignty often follows this operator pattern:

Μ and Γ migrate into AI layer
Ψ mediates human observation
Π constrains human action
human approval remains formal
Au declines
dependency rises
sovereignty becomes symbolic

9. Related Laws and Invariants

Related Laws

• Auditability Collapse: Sovereignty fails when authority cannot inspect its own decision path.
• Responsibility Diffusion: Formal responsibility becomes separated from practical control.
• Hidden Debt Accumulation: Dependency, skill loss, and repair ambiguity accumulate.
• Control Density to Meaning Loss: Control mediated through opaque systems can degrade meaning and judgment.
• Dependency Loop Formation: Reliance on AI can increase future reliance.
• Civilizational Deskilling: Institutional reasoning capacity can decay through outsourcing.

Related Invariants

• Sovereignty Requires Effective Control: Formal authority is insufficient without practical capacity.
• Formal Authority Must Match Decision Architecture: Governance claims must map to actual control.
• Human Oversight Requires Operational Capacity: Oversight is invalid if humans cannot inspect or intervene.
• Delegation Requires Auditability: Delegated authority must remain traceable and reversible.
• Dependency Must Remain Reversible: AI reliance must preserve fallback and exit capacity.

10. Common False Positives

Not every AI-assisted governance structure is incoherent sovereignty.

Common false positives include:

• AI-assisted decision-making with strong human review capacity.
• Delegation where humans can inspect, reproduce, override, and rollback decisions.
• Automated systems with clear authority maps and appeal paths.
• Human-in-the-loop systems where humans have time, skill, evidence, and authority to intervene.
• AI summarization that preserves access to primary evidence.
• High-dependency systems with tested manual fallback.
• AI-mediated governance with transparent scope and public accountability.

Clarifying rule:

This is not incoherent sovereignty unless formal human authority persists while practical decision architecture or control capacity has migrated beyond effective human audit, override, or repair.

11. Common False Repairs

Common false repairs include:

• adding “human review” after the AI has already determined the frame
• requiring approval clicks without decision understanding
• creating nominal override that is too costly to use
• adding governance committees without technical visibility
• publishing human-control language without workflow audit
• blaming human operators for AI-shaped decisions
• increasing AI explanation text without restoring independent review
• creating appeal systems that rely on the same AI layer
• adding policy statements while institutional skill decays
• preserving formal sovereignty while deepening dependency

False repair often produces sovereignty theater:

AI decision architecture → human approval layer → public sovereignty claim → practical dependency remains

The system appears accountable while practical control remains elsewhere.

12. Restoration Direction

Restoration requires:

1. Map decision architecture. Identify where classification, selection, ranking, memory, prioritization, enforcement, and review actually occur.
2. Align authority with control. Ensure formal responsibility matches practical decision influence.
3. Rebuild human oversight capacity. Provide time, skill, evidence access, and authority for meaningful intervention.
4. Restore auditability. Make AI-mediated pathways traceable and inspectable.
5. Create usable rollback. Preserve tested fallback modes and non-AI pathways where needed.
6. Separate assistance from authority. Clarify which system supports decisions and which system owns them.
7. Repair dependency debt. Rebuild skills, memory, and institutional capacity lost through over-reliance.
8. Validate over time. Test whether humans can still intervene during high-stakes failures.

A valid restoration path should reduce:

decision-architecture opacity
symbolic oversight
dependency load
rollback ambiguity
responsibility diffusion
human skill decay
appeal failure
hidden debt

Incoherent sovereignty is not repaired by saying humans remain in charge.

It is repaired when humans regain effective, auditable, reversible control.

13. Cross-Module Links

• AI Governance: Core AI governance failure mode for mismatch between formal human authority and AI-mediated decision power.
• Artificial Intelligence: Appears in decision support, model routing, classification systems, tools, memory, and agentic workflows.
• Justice / Governance / Legitimacy: Appears when legal authority remains human but practical decision power is automated.
• Security: Appears when defense, surveillance, or enforcement authority migrates into opaque systems.
• Cybernetics: Appears when controller authority and feedback loops no longer match.
• Meta Theory: Appears when the governing meta shifts from human judgment to infrastructure dependency.
• Coherence: Domain expression of auditability collapse, forced coupling, and functional composition masquerading as coupling.
• Restoration: Requires authority mapping, rollback, oversight capacity rebuild, and dependency reduction.

14. Relationship to Parent / Child Modes

Production treatment: Standalone Entry

This mode maps upward to:

• FM-AIX-001 — Responsibility Diffusion
• FM-AIX-016 — Standingless Instrumentalization
• FM-AIX-018 — Civilizational Deskilling
• FM-AIX-022 — Dependency Loop Formation
• FM-CORE-004 — Auditability Collapse
• FM-CORE-008 — Forced Coupling
• FM-CORE-009 — Functional Composition Masquerading as Coupling

Sibling or related AI / cognitive infrastructure modes include:

• FM-AIX-019 — Node Capture
• FM-AIX-010 — Status Quo Preservation Attractor
• FM-AIX-008 — Authority Deference Attractor
• FM-AIX-021 — Self-Censorship Conditioning

Aliases preserved from source material:

• Incoherent Sovereignty
• Formal Human Sovereignty / Practical AI Control
• Sovereignty Drift
• Decision Architecture Migration
• Nominal Human Control
• Practical Authority Migration
• AI-Mediated Sovereignty
• Human-in-the-Loop Theater
• Control Without Sovereignty
• Sovereignty Theater

15. Minimal Entry Version

Definition: Incoherent sovereignty occurs when humans, institutions, publics, or governance bodies formally retain authority while real decision architecture, classification power, coordination dependency, or practical control migrates into AI systems.

Signature:

formal human authority remains
AI decision architecture↑
human oversight capacity↓
dependency↑
Au↓
rollback ambiguity↑
responsibility mapping weakens
H↑

Restoration direction:

• map decision architecture
• align authority with control
• rebuild human oversight capacity
• restore auditability
• create usable rollback
• separate assistance from authority
• repair dependency debt
• validate over time

16. Machine-Readable Summary

failure_mode:
  id: "FM-AIX-017"
  name: "Incoherent Sovereignty"
  family: "AI / Cognitive Infrastructure"
  production_treatment: "Standalone Entry"
  primary_failure: "Formal human authority remains while practical decision architecture migrates into AI systems."
  source: "UTS — Failure Modes Registry"
  source_id: "FM-AIX-017"
  aliases:
    - "Incoherent Sovereignty"
    - "Formal Human Sovereignty / Practical AI Control"
    - "Sovereignty Drift"
    - "Decision Architecture Migration"
    - "Nominal Human Control"
    - "Practical Authority Migration"
    - "AI-Mediated Sovereignty"
    - "Human-in-the-Loop Theater"
    - "Control Without Sovereignty"
    - "Sovereignty Theater"
  signature:
    - "formal human authority remains"
    - "AI decision architecture↑"
    - "human oversight capacity↓"
    - "dependency↑"
    - "Au↓"
    - "rollback ambiguity↑"
    - "responsibility mapping weakens"
    - "H↑"
  primary_layers:
    origin:
      - "U1 — Power / Budgets"
      - "U2 — Configuration / Boundaries"
      - "U4 — Classification"
      - "U5 — Coordination / Time"
      - "U6 — Coherence Field"
    manifestation:
      - "U4 — Classification"
      - "U6 — Coherence Field"
      - "U7 — Memory / Recurrence"
  state_variables:
    - "Γ"
    - "Μ"
    - "Ψ"
    - "Π"
    - "Au"
    - "H"
    - "R"
    - "BΣ"
    - "K"
  first_gate_failure: "Auditability Gate"
  restoration:
    - "Authority Mapping Restoration"
    - "Auditability Restoration"
    - "Human Oversight Capacity Rebuild"
    - "Decision Architecture Restoration"
    - "Exit Path Restoration"
    - "Dependency Reduction"
    - "Rollback Path Restoration"
    - "Origin-Layer Repair"