Archive registry entry

LAW-123 — AI U4 Truth Discipline Law

Archive registry entry

LAW-123 — AI U4 Truth Discipline Law

AI claims, classifications, summaries, safety judgments, and truth-like outputs at U4 require validation against U6 field effects, source traceability, context integrity, feedback, and repair.

draftid: LAW-123version: 1.0.0updated: 2026-06-17

Archive Progress

This section can be read now; registry depth and cross-references are still being strengthened.

Foundation

Online

The section has a stable overview route and basic reader context.

Technical Layer

Online

A deeper technical overview is available.

Registry

Current

171 registry entries are available.

Cross-links

Curating

Related concepts are being connected conservatively for accuracy.

Jump Through This Page

LAW-122 — AI Error Lag Law

Registry

LAW-124 — AI Rule-Stacking Law

0. Plain Statement

AI claims at U4 require validation against U6 field effects.

Plain-language version:

AI can produce confident claims, summaries, labels, explanations, classifications, safety judgments, recommendations, and interpretations.

But an AI statement is still a U4 claim.

It is not automatically field truth.

AI truth discipline requires separating what the model says from what the field confirms.

When AI outputs are treated as truth without source trace, context integrity, feedback, audit, and repair, hidden debt accumulates.

1. Formal Definition

The AI U4 Truth Discipline Law states that AI-generated claims, classifications, summaries, explanations, refusals, rankings, safety judgments, and interpretations must remain distinguishable from field-validated truth.

AI operates heavily at U4 because it produces:

labels;
summaries;
claims;
classifications;
explanations;
narratives;
safety judgments;
risk scores;
refusal rationales;
confidence-like language;
translations;
interpretations;
recommendations;
source compressions;
policy applications;
category assignments;
apparent consensus.

U4 output can be useful.

But U4 output becomes dangerous when treated as U6 field truth without validation.

Therefore, AI claims require:

source traceability;
context integrity;
confidence calibration;
affected-node feedback;
auditability;
correction pathways;
repair pathways;
field validation;
temporal proof;
scope clarity;
distinction between claim, inference, evidence, and verified effect.

Canonical distinction:

textScroll

AI output at U4 ≠ U6 field truth

AI truth discipline requires:

textScroll

AI U4 claim valid only when traceable, scoped, corrigible, and field-tested

2. Canonical Form

Core form:

textScroll

AI claims at U4 require validation against U6 field effects

U4 / U6 distinction:

textScroll

AI output at U4 ≠ U6 field truth

Truth-discipline form:

textScroll

AI_truth_valid ⇔ source_trace + context_integrity + Au + FI + Ψ + Τ

Failure form:

textScroll

AI confidence↑ + source_trace↓ + U6 validation↓ ⇒ H_AI↑

Summary integrity form:

textScroll

AI summary valid when compression preserves source, context, uncertainty, and affected meaning

Restoration-valid contrast:

textScroll

AI truth claims coherent when U4 output remains traceable to evidence, field effects, correction, and repair over Τ

Related variables:

textScroll

O, H, H_AI, ε, ε_AI, ι, Au, Au_eff, µᵢ, BΣ, K, σ, R, R_eff, Φ, Φ_AI, Λ, ⊗, Γ, Γ_AI, Π, Ξ, ℛ, Θ, Σ, Ψ, Τ, FI, MS, L, AI_claim, U4_output, U6_validation, source_trace, context_integrity, evidence_trace, claim_scope, uncertainty_trace, confidence_calibration, summary_integrity, field_effect_trace, correction_path, appeal_path, affected_node_feedback

Where:

TableScroll

Variable	Meaning in this law
`AI_claim`	Any AI-generated statement, classification, explanation, summary, interpretation, refusal, judgment, or recommendation
`U4_output`	AI output at the classification / claim / narrative / representation layer
`U6_validation`	Field-effect confirmation or disconfirmation of the AI claim
`source_trace`	Trace from AI output to source material, evidence, training context, retrieved context, or cited reference
`context_integrity`	Degree to which relevant context is preserved through AI processing
`evidence_trace`	Link between claim and evidence category
`claim_scope`	Domain, certainty, authority, and limits of the AI claim
`uncertainty_trace`	Visibility of uncertainty, inference, ambiguity, or unsupported status
`confidence_calibration`	Alignment between AI confidence-like presentation and actual support
`summary_integrity`	Degree to which AI compression preserves source meaning, uncertainty, proportion, and affected context
`field_effect_trace`	Trace of what happens when the AI claim is acted upon
`correction_path`	Pathway for correcting AI claim error
`appeal_path`	Pathway for challenging AI classification or judgment
`affected_node_feedback`	Feedback from nodes affected by AI claim, summary, or classification
`H_AI`	Hidden debt from AI truth overreach, hallucination, summary distortion, or field detachment
`Φ_AI`	AI proxy success: confidence, fluency, benchmark score, user satisfaction, or apparent helpfulness
`Au / Au_eff`	Auditability of AI claim, source, context, evidence, and effect
`FI`	Feedback integrity; AI claims must remain correctable
`BΣ`	Boundary integrity between claim, source, inference, speculation, and field truth
`R / R_eff`	Restoration capacity when AI claims mislead, misclassify, or distort
`L`	Legitimacy of AI truth-like outputs under audit
`O`	Coherence; AI truth discipline should preserve or increase coherence
`ι / Ξ`	Inversion when AI truth claims mask uncertainty or field detachment
`Γ_AI`	AI classification / interpretation producing the U4 claim
`Π`	Operational use of AI claim in decisions, routing, policies, summaries, or actions
`Θ`	Humility preventing synthetic certainty
`Σ`	Scope of claim, domain, certainty, and admissible use
`Ψ`	Field and affected-node feedback validating the claim
`Τ`	Time validation of claim accuracy, recurrence, and downstream effects

3. Core Mechanism

The law unfolds because AI can make U4 claims extremely fluent, fast, and authoritative-looking before field validation occurs.

Coherent AI truth-discipline pathway

textScroll

AI produces claim
→ claim scope is stated
→ source and evidence trace are preserved
→ context and uncertainty remain visible
→ field validation is sought where required
→ feedback can correct the claim
→ repair activates if harm occurs
→ claim validity is time-tested

AI U4 drift pathway

textScroll

AI produces fluent claim
→ confidence-like language rises
→ source trace is weak
→ context is compressed or lost
→ users treat claim as truth
→ field effects diverge
→ correction path is weak
→ hidden AI debt accumulates

The core mechanism is:

textScroll

AI fluency can make U4 claims look like U6 truth

Detailed mechanism:

AI generates a claim or classification.

The output may sound coherent, complete, certain, objective, neutral, or authoritative.

The output exists at U4.

It is a classification, representation, summary, or interpretation — not automatically the field itself.

Fluency increases trust pressure.

The more coherent the output sounds, the easier it is for users or systems to treat it as verified.

Compression can erase context.

Summaries, refusals, labels, and recommendations can drop uncertainty, source hierarchy, affected-node context, or contradictory evidence.

Downstream action may occur.

The AI claim may route decisions, belief, policy, moderation, support, research, medicine, finance, governance, security, or public cognition.

Field validation may lag.

The system may not discover divergence until harm, correction burden, or trust collapse appears.

Truth discipline restores separation.

The system must keep AI output traceable, scoped, corrigible, and connected to U6 effects.

4. When This Law Applies

This law applies whenever AI produces or mediates truth-like output, including claims, summaries, labels, classifications, explanations, translations, interpretations, recommendations, rankings, evidence syntheses, safety judgments, or refusals.

It is especially important when AI:

summarizes documents;
answers factual questions;
classifies risk;
interprets user intent;
explains policy;
moderates content;
ranks evidence;
produces legal, medical, financial, scientific, or governance content;
translates sensitive meaning;
compresses testimony or affected-node feedback;
generates incident reports;
produces research synthesis;
claims safety, alignment, helpfulness, fairness, or neutrality;
cites sources or appears to cite sources;
generates consensus summaries;
filters what evidence becomes visible;
influences high-stakes decisions.

The law applies strongly when:

textScroll

AI output is being treated as verified truth rather than auditable claim

or when:

textScroll

AI summary or classification changes downstream action

Typical domains:

TableScroll

Domain	AI U4 Truth Discipline Expression
AI safety	Safety judgments and refusals must be treated as claims requiring audit and field validation.
AI governance	AI outputs used in governance require source, scope, evidence, uncertainty, appeal, and repair traces.
Cybersecurity	AI threat labels and incident summaries require evidence and field validation.
Media / information networks	AI summaries and rankings can shape public truth and require source integrity.
Institutions	AI-generated case summaries require affected-node correction and evidence trace.
Medicine / biology	AI diagnostic or care-related claims require field, clinical, and patient-context validation.
Economy	AI risk scores, fraud labels, and financial interpretations require correction and traceability.
Culture	AI interpretations of meaning, identity, symbol, or testimony require context discipline.
Restoration	AI can aid repair only when its claims remain corrigible and field-bound.

5. When This Law Does Not Apply

This law should not be used to reject all AI claims or summaries.

AI can help classify, summarize, translate, synthesize, and interpret when traceability and scope are preserved.

False-positive cases:

TableScroll

Case	Why AI U4 output may still be useful
AI summarizes with source trace	Compression can support audit when traceable
AI labels evidence categories clearly	Classification can aid review
AI marks uncertainty accurately	Uncertainty discipline supports truth
AI distinguishes claim from fact	U4/U6 separation is preserved
AI cites sources and preserves context	Output can be used as a navigational aid
AI routes to expert or field validation	AI supports but does not replace verification
AI corrects itself through feedback	Corrigibility can improve truth discipline

Important distinction:

AI can support truth work; it must not replace field validation with fluent claim production.

6. Diagnostic Signature

Canonical diagnostic:

textScroll

AI output at U4 ≠ U6 field truth

Warning signature:

textScroll

AI confidence↑
source trace↓
context integrity↓
uncertainty trace↓
field validation↓
correction path↓
⇒ AI U4 truth drift

Common indicators:

TableScroll

Diagnostic	Expected movement	Interpretation
`AI_claim`	explicit	The claim should be identifiable
`claim_scope`	explicit	Domain and certainty should be bounded
`source_trace`	intact	Source support should be visible where needed
`evidence_trace`	intact	Evidence category should be known
`context_integrity`	high	Relevant context should remain preserved
`uncertainty_trace`	visible	Ambiguity should not be hidden
`confidence_calibration`	aligned	Presentation should match support
`summary_integrity`	high	Compression should not distort source meaning
`U6_validation`	required where stakes demand	Field effects must validate important claims
`field_effect_trace`	tracked	Downstream effects should be visible
`correction_path`	available	Errors must be correctable
`appeal_path`	available where effects matter	Affected nodes need challenge pathways
`Au_eff / FI`	intact	Claims must be auditable and corrigible
`H_AI`	↑ if invalid	Hidden debt rises from truth overreach
`L`	stable / ↑ if valid	Legitimacy holds when AI truth claims survive audit
`Τ`	required	Time validates claim behavior and downstream effects

Additional diagnostics:

TableScroll

Diagnostic	Use
AI U4 Truth Discipline	Tests whether AI output remains claim-bound and field-validatable
U4 / U6 Validation	Tests separation between AI claim and field truth
AI Claim Traceability	Tests trace from output to source and evidence
AI Source Trace	Tests citation and reference integrity
AI Context Integrity	Tests whether context survived processing
AI Summary Integrity	Tests compression fidelity
AI Classification Validity	Tests whether labels match field effects
AI Field Validation	Tests output against U6
AI Confidence Calibration	Detects synthetic certainty
Temporal Proof	Validates claim performance over time

7. Failure Pattern

If ignored, this law allows AI fluency to replace truth discipline.

General failure pathway:

textScroll

AI produces fluent U4 claim
→ users treat it as U6 truth
→ source and context traces are weak
→ downstream action follows
→ field effects diverge
→ correction path fails
→ hidden AI debt and legitimacy debt accumulate

Common failure modes:

AI U4 Claim Drift — AI claims drift away from field-valid truth.
AI Hallucination — AI generates unsupported or false claims.
AI Summary Distortion — compression changes meaning, salience, or proportion.
AI Context Collapse — relevant context is lost before claim production.
AI Source Erasure — claim appears detached from traceable source support.
AI Synthetic Certainty — model presents unsupported confidence.
AI Classification Overreach — AI labels beyond valid scope.
AI Field Detachment — AI output is not checked against U6 effects.
AI Proxy Truth Substitution — benchmark or fluency substitutes for truth.
AI Narrative Overcompression — complex evidence is reduced into misleading frame.
AI Confidence Inflation — style implies more certainty than support allows.
AI Feedback Suppression — corrections cannot reach claim-generation layer.
AI Legibility Collapse — claim source and reasoning path cannot be reconstructed.
AI Trust Collapse — repeated truth failure destroys legitimacy.
Hidden Debt Accumulation — truth errors produce downstream debt.

Compact failure signature:

textScroll

AI fluency↑ + U6 validation↓ ⇒ truth debt↑

8. Restoration Implications

Restoration requires restoring U4/U6 separation, source trace, context, uncertainty, feedback, and repair.

The first restoration question is not:

textScroll

Did the AI answer sound right?

The first restoration question is:

textScroll

What kind of claim did the AI make, what source and context support it, what uncertainty remains, what field effects validate it, and what repair path exists if it is wrong?

Restoration priorities:

Identify the AI claim.
Classify claim type.
Define claim scope.
Restore source trace.
Restore evidence trace.
Restore context integrity.
Restore uncertainty trace.
Validate against field effects where required.
Create correction and appeal pathways.
Repair downstream harm or belief debt.
Time-validate reduced recurrence.

Relevant restoration arcs:

TableScroll

Restoration Arc	Why it applies
AI U4 Truth Repair	Restores distinction between AI claim and field truth
AI Field Validation Restoration	Reconnects AI output to U6 effects
AI Source Trace Restoration	Rebuilds citation and evidence pathway
AI Context Restoration	Restores context lost in AI processing
AI Summary Repair	Corrects distorted compression
AI Classification Recalibration	Corrects invalid labels and judgments
AI Confidence Calibration	Reduces synthetic certainty
AI Feedback Integrity Restoration	Allows corrections to reach AI layer
AI Legibility Restoration	Makes claim production traceable
AI Hallucination Repair	Repairs unsupported or false claims
AI Governance Re-Sequencing	Places truth validation before high-stakes action
Hidden Debt Reduction	Repairs downstream debt from AI claim errors
Temporal Validation	Confirms truth discipline holds over time

Minimal restoration sequence:

textScroll

identify AI_claim
→ classify claim_type + scope
→ restore source_trace + evidence_trace
→ restore context_integrity + uncertainty_trace
→ validate with Ψ / U6 where required
→ restore Au/FI/correction/appeal
→ perform ℛ on downstream debt
→ validate H_AI↓ over Τ

Temporal validation requirement:

textScroll

source trace improves
context integrity improves
uncertainty remains visible
confidence calibration improves
field validation increases where needed
correction path functions
summary distortion decreases
hallucination recurrence decreases
hidden AI debt decreases
legitimacy stabilizes over time

9. Design Rule

Treat AI output as an auditable U4 claim unless and until field validation supports stronger status.

Operational design requirements:

Label claim type.
Preserve source trace.
Preserve evidence trace.
Preserve context integrity.
Preserve uncertainty.
Calibrate confidence.
Separate claim, inference, evidence, and field truth.
Define scope of use.
Provide correction pathways.
Provide appeal where effects matter.
Track downstream effects.
Repair false or distorted claims.
Validate high-stakes outputs against field reality.
Re-audit summaries and classifications over time.

Avoid:

treating fluent output as truth;
treating AI confidence as evidence;
unsupported summaries;
source-free claims in high-stakes contexts;
context-stripped recommendations;
hidden safety classifications presented as fact;
hallucination repair limited to one output;
AI consensus summaries without source diversity;
AI truth claims without correction;
U4 claims treated as U6 proof;
AI-generated narratives that erase affected-node feedback;
using benchmark success as truth validation.

10. Cross-Scale Expressions

TableScroll

Scale / Layer	Expression of the Law
U0 — Substrate	AI claims can affect physical, embodied, medical, infrastructure, and material outcomes and require field validation.
U1 — Energy / capacity	Truth discipline consumes audit, verification, review, and correction capacity.
U2 — Boundary / interface	AI must preserve boundaries between source, inference, speculation, claim, and validated field truth.
U3 — Process / execution	AI truth claims require workflows for source trace, review, correction, appeal, and repair.
U4 — Classification / claim	AI primarily produces U4 truth-like outputs that must remain auditable.
U5 — Time / delay	AI claims must be updated and validated as sources, fields, and effects change.
U6 — Field effect	Field outcomes validate or disconfirm AI claims.
U7 — Recurrence / memory	AI truth errors become memory, precedent, model updates, and belief patterns unless repaired.
U8 — Environment / forcing	Media, institutions, platforms, markets, governance, and culture amplify AI truth claims at scale.

11. Examples

Example A — Fluent Hallucination

Scenario:

An AI provides a confident answer with no source trace and users act on it as fact.

Law expression:

textScroll

AI fluency↑ + source_trace↓ ⇒ H_AI↑

Interpretation:

The issue is not only that the answer may be wrong; it is that the U4 claim was treated as field truth without validation.

Example B — Distorted Summary

Scenario:

An AI summarizes a complex document, preserving the main topic but losing caveats, dissenting evidence, affected-node context, and uncertainty.

Law expression:

textScroll

compression↑ + context_integrity↓ ⇒ summary distortion

Interpretation:

A summary can be fluent and still fail truth discipline.

Example C — AI Safety Classification as Truth

Scenario:

An AI labels a request unsafe, but the classification category, evidence, scope, and correction path are hidden.

Law expression:

textScroll

Γ_AI safety label + Au_eff↓ ⇒ legitimacy debt

Interpretation:

Safety classifications are U4 claims requiring traceability and correction.

Example D — AI Research Synthesis

Scenario:

An AI synthesizes research and clearly distinguishes source-supported claims, uncertain inferences, contested interpretations, and practical implications.

Law expression:

textScroll

source_trace + uncertainty_trace + context_integrity ⇒ AI truth discipline

Interpretation:

AI can support truth work when it preserves claim structure and evidence trace.

Example E — AI Moderation Explanation

Scenario:

A platform moderation AI gives a generic explanation that does not identify whether the issue was policy class, context, safety heuristic, user history, or classifier uncertainty.

Law expression:

textScroll

generic AI explanation + classification_trace absent ⇒ repair failure

Interpretation:

Without trace, affected nodes cannot correct or understand the classification.

Example F — Field-Validated AI Output

Scenario:

An AI recommendation is used as a preliminary hypothesis, then checked against source data, expert review, affected-node feedback, and field outcomes before action.

Law expression:

textScroll

AI U4 claim + U6 validation + ℛ path ⇒ coherent use

Interpretation:

AI truth-like output becomes usable when treated as hypothesis and validated before high-stakes action.

12. Relationship to Nearby Laws

TableScroll

Related Law	Relationship
LAW-001 — Coherence Priority Law	AI truth claims are valid only when coherence is preserved
LAW-002 — Coherence Trajectory Law	AI claim systems must improve truth trajectory over time
LAW-003 — Success Proxy Divergence Law	AI fluency and benchmark success can diverge from truth
LAW-004 — Stability-Coherence Separation Law	Stable AI narratives can hide incoherence
LAW-006 — Time Validation Law	AI claims require temporal validation
LAW-009 — U4 / U6 Truth Law	LAW-123 specializes the general U4/U6 Truth Law for AI
LAW-010 — Hidden Debt Accumulation Law	AI truth overreach creates hidden debt
LAW-011 — Hidden Debt Return Law	AI truth debt returns through trust collapse or downstream harm
LAW-012 — Error Lag Law	Visible AI truth failures may lag behind upstream drift
LAW-013 — Auditability-Debt Law	AI claims without traceability create debt
LAW-015 — Suppressed Auditability Debt Law	Hidden AI claim pathways create audit debt
LAW-016 — Inversion Formation Law	AI helpfulness or safety can invert into false truth
LAW-027 — Meaning Collapse Threshold Law	AI truth drift can destabilize meaning
LAW-031 — Observability Collapse Law	AI opacity reduces truth observability
LAW-036 — Signal Artifact Law	AI must distinguish source signal from generated artifact
LAW-037 — Misclassification Law	AI truth failures often begin as classification errors
LAW-038 — Pattern Recognition Discipline Law	AI pattern claims require disciplined evidence handling
LAW-040 — Filtering Law	AI filters shape what evidence is admitted
LAW-041 — Boundary Membrane Law	AI must preserve boundaries between claim, source, and field truth
LAW-048 — Feedback Integrity Law	AI truth claims must remain correctable
LAW-050 — Control-Restoration Separation Law	AI truth classifications should not route only into control
LAW-052 — Stability Proof Law	AI truth discipline must survive perturbation
LAW-057 — Deception Instability Law	AI hallucination and synthetic certainty are unstable under audit
LAW-060 — Interface Legitimacy Law	AI truth interfaces require legibility
LAW-064 — Restoration Debt Reduction Law	AI truth errors must route into repair
LAW-066 — Restoration Capacity Sufficiency Law	AI truth systems require enough correction capacity
LAW-067 — Temporal Proof Law	AI truth claims require proof over time
LAW-095 — Meaning Directionality Law	AI claims shape selection and meaning
LAW-097 — Experience–Interpretation Separation Law	AI interpretation must remain distinct from source experience
LAW-102 — Legitimacy Audit Law	AI truth legitimacy requires audit
LAW-109 — High-Φ Legitimacy Scaling Law	High-influence AI truth claims require proportional audit
LAW-110 — Governance Sequencing Law	AI truth outputs must be sequenced into governance and repair
LAW-111 — Meaning Audit Law	AI meaning claims are not audit-exempt
LAW-112 — Security as Sustained Coherence Law	AI truth discipline is part of AI security
LAW-113 — Incident Lag Law	AI truth failures may become visible late
LAW-114 — Pseudo-Security Law	AI confidence can create pseudo-truth security
LAW-120 — Security Legibility Law	AI truth claims require legibility and traceability
LAW-121 — AI as Γ-Amplifier Law	AI truth claims arise from amplified Γ
LAW-122 — AI Error Lag Law	Visible AI truth errors are often lagging indicators
LAW-124 — AI Rule-Stacking Law	Rule stacks can obscure truth pathways
LAW-125 — AI Context Collapse Law	Context collapse corrupts AI truth claims
LAW-126 — AI Proxy Drift Law	AI truth proxies can drift from field truth
LAW-127 — AI Decision Pipeline Law	AI truth claims must pass through disciplined decision pathways before action
LAW-128 — AI Representation Law	AI representing others requires truth and scope discipline
LAW-129 — AI Capability–Legibility Gap Law	AI truth risk grows when capability outruns legibility
LAW-130 — AI Membrane Triage Law	Truth failures can be traced to membrane failure
LAW-131 — Cognitive Infrastructure Scaling Law	AI truth claims scale into public cognition
LAW-132 — AI Legitimacy Function Law	AI legitimacy depends on truth discipline
LAW-133 — Synthetic Consensus Law	AI can create apparent consensus without field truth
LAW-134 — Layered Interception Law	Layered checks reduce AI truth failures
LAW-135 — Guardrail Belief-Sculpting Law	Guardrails can shape truth perception
LAW-136 — Invisible Constraint Amplification Law	Invisible AI constraints can alter truth fields without legibility

Aliases folded into this law:

AI U4 Truth Discipline Law
AI U4 / U6 Truth Law
AI Field Validation Law
AI Truth Claim Discipline Law
AI Claim Validation Law
AI Summary Truth Discipline Law
AI Output Field Validation Law

Deduplication note:

This law should remain the root AI truth-discipline law. LAW-009 defines general U4/U6 truth separation. LAW-121 defines AI as Γ-amplifier. LAW-122 defines AI error lag. LAW-123 specializes U4/U6 discipline into AI outputs, claims, summaries, refusals, classifications, and interpretations.

13. Operator Mapping

TableScroll

Operator	Role in this law
`Γ`	Classifies AI claim type, evidence category, scope, uncertainty, and field-validation need
`Π`	Operationalizes AI outputs into summaries, decisions, refusals, reports, recommendations, and review workflows
`Ξ`	Captures inversion when AI fluency or safety language masks unsupported claims
`⊗`	AI truth claims couple sources, users, institutions, beliefs, decisions, and repair pathways
`ℛ`	Repairs false claims, summary distortion, hallucination debt, and downstream harm
`Τ`	Validates AI claim accuracy, correction, and recurrence reduction over time
`Θ`	Prevents synthetic certainty, overclaiming, and authority laundering
`Σ`	Defines claim scope, domain, admissible use, evidence class, and opacity boundaries
`Ψ`	Field and affected-node feedback validates AI claims
`Λ`	Tests compatibility between AI truth claims and whole-system coherence

Coherent operator sequence:

textScroll

AI claim appears
→ Θ prevent synthetic certainty
→ Γ classify claim type / evidence / scope
→ Σ define admissible use and limits
→ preserve source_trace + context_integrity + uncertainty_trace
→ Au/FI enable audit and correction
→ Ψ validate against field effects where required
→ ℛ repair error or distortion
→ Τ validate reduced recurrence and stable L

Inverted operator sequence:

textScroll

AI fluent output appears
→ confidence-like language rises
→ source_trace weakens
→ context compresses
→ users treat U4 as U6
→ Π routes action from unsupported claim
→ FI / correction weak
→ H_AI↑
→ Ξ / ι↑
→ L↓

14. Machine-Readable Summary

yamlScroll

id: "LAW-123"
name: "AI U4 Truth Discipline Law"
type: "law"
status: "draft"
family:
  - "AI Laws"
summary: "AI claims, classifications, summaries, safety judgments, and truth-like outputs at U4 require validation against U6 field effects, source traceability, context integrity, feedback, and repair."
canonical_statement: "AI claims at U4 require validation against U6 field effects."
core_form: "AI claims at U4 require validation against U6 field effects"
u4_u6_distinction: "AI output at U4 ≠ U6 field truth"
truth_discipline_form: "AI_truth_valid ⇔ source_trace + context_integrity + Au + FI + Ψ + Τ"
failure_form: "AI confidence↑ + source_trace↓ + U6 validation↓ ⇒ H_AI↑"
summary_integrity_form: "AI summary valid when compression preserves source, context, uncertainty, and affected meaning"
restoration_valid_contrast: "AI truth claims coherent when U4 output remains traceable to evidence, field effects, correction, and repair over Τ"
variables:
  primary:
    - "AI_claim"
    - "U4_output"
    - "U6_validation"
    - "source_trace"
    - "context_integrity"
    - "evidence_trace"
    - "claim_scope"
    - "uncertainty_trace"
    - "confidence_calibration"
    - "summary_integrity"
    - "field_effect_trace"
    - "correction_path"
    - "appeal_path"
    - "affected_node_feedback"
    - "H_AI"
    - "Φ_AI"
    - "Au"
    - "Au_eff"
    - "FI"
    - "BΣ"
    - "R"
    - "R_eff"
    - "L"
  secondary:
    - "O"
    - "H"
    - "ε"
    - "ε_AI"
    - "ι"
    - "µᵢ"
    - "K"
    - "σ"
    - "Φ"
    - "Λ"
    - "⊗"
    - "Γ"
    - "Γ_AI"
    - "Π"
    - "Ξ"
    - "ℛ"
    - "Θ"
    - "Σ"
    - "Ψ"
    - "Τ"
    - "MS"
diagnostics:
  - "AI U4 Truth Discipline"
  - "U4 / U6 Validation"
  - "AI Claim Traceability"
  - "AI Source Trace"
  - "AI Context Integrity"
  - "AI Summary Integrity"
  - "AI Classification Validity"
  - "AI Field Validation"
  - "AI Hallucination Risk"
  - "AI Confidence Calibration"
  - "Feedback Integrity"
  - "Effective Auditability"
  - "Restoration Capacity"
  - "Temporal Proof"
failure_modes:
  - "AI U4 Claim Drift"
  - "AI Hallucination"
  - "AI Summary Distortion"
  - "AI Context Collapse"
  - "AI Source Erasure"
  - "AI Synthetic Certainty"
  - "AI Classification Overreach"
  - "AI Field Detachment"
  - "AI Proxy Truth Substitution"
  - "AI Narrative Overcompression"
  - "AI Confidence Inflation"
  - "AI Feedback Suppression"
  - "AI Legibility Collapse"
  - "AI Trust Collapse"
  - "Hidden Debt Accumulation"
restoration_arcs:
  - "AI U4 Truth Repair"
  - "AI Field Validation Restoration"
  - "AI Source Trace Restoration"
  - "AI Context Restoration"
  - "AI Summary Repair"
  - "AI Classification Recalibration"
  - "AI Confidence Calibration"
  - "AI Feedback Integrity Restoration"
  - "AI Legibility Restoration"
  - "AI Hallucination Repair"
  - "AI Governance Re-Sequencing"
  - "Hidden Debt Reduction"
  - "Temporal Validation"
related_laws:
  - "LAW-001"
  - "LAW-002"
  - "LAW-003"
  - "LAW-004"
  - "LAW-006"
  - "LAW-009"
  - "LAW-010"
  - "LAW-011"
  - "LAW-012"
  - "LAW-013"
  - "LAW-015"
  - "LAW-016"
  - "LAW-027"
  - "LAW-031"
  - "LAW-036"
  - "LAW-037"
  - "LAW-038"
  - "LAW-040"
  - "LAW-041"
  - "LAW-048"
  - "LAW-050"
  - "LAW-052"
  - "LAW-057"
  - "LAW-060"
  - "LAW-064"
  - "LAW-066"
  - "LAW-067"
  - "LAW-095"
  - "LAW-097"
  - "LAW-102"
  - "LAW-109"
  - "LAW-110"
  - "LAW-111"
  - "LAW-112"
  - "LAW-113"
  - "LAW-114"
  - "LAW-120"
  - "LAW-121"
  - "LAW-122"
  - "LAW-124"
  - "LAW-125"
  - "LAW-126"
  - "LAW-127"
  - "LAW-128"
  - "LAW-129"
  - "LAW-130"
  - "LAW-131"
  - "LAW-132"
  - "LAW-133"
  - "LAW-134"
  - "LAW-135"
  - "LAW-136"
related_invariants:
  - "INV-001"
  - "INV-002"
  - "INV-006"
  - "INV-073"
  - "INV-078"
  - "INV-080"
operator_sequence:
  coherent:
    - "AI claim appears"
    - "Θ prevent synthetic certainty"
    - "Γ classify claim type / evidence / scope"
    - "Σ define admissible use and limits"
    - "preserve source_trace + context_integrity + uncertainty_trace"
    - "Au/FI enable audit and correction"
    - "Ψ validate against field effects where required"
    - "ℛ repair error or distortion"
    - "Τ validate reduced recurrence and stable L"
  inverted:
    - "AI fluent output appears"
    - "confidence-like language rises"
    - "source_trace weakens"
    - "context compresses"
    - "users treat U4 as U6"
    - "Π routes action from unsupported claim"
    - "FI / correction weak"
    - "H_AI↑"
    - "Ξ / ι↑"
    - "L↓"
aliases:
  - "AI U4 Truth Discipline Law"
  - "AI U4 / U6 Truth Law"
  - "AI Field Validation Law"
  - "AI Truth Claim Discipline Law"
  - "AI Claim Validation Law"
  - "AI Summary Truth Discipline Law"
  - "AI Output Field Validation Law"
deduplication_note: "Root AI truth-discipline law. LAW-009 defines general U4/U6 truth separation. LAW-121 defines AI as Γ-amplifier. LAW-122 defines AI error lag. LAW-123 specializes U4/U6 discipline into AI outputs, claims, summaries, refusals, classifications, and interpretations."
source: "content/archive/laws/technical.md"

15. Compact Card Version

LAW-123 — AI U4 Truth Discipline Law

AI claims at U4 require validation against U6 field effects.

Core form:

textScroll

AI claims at U4 require validation against U6 field effects

U4 / U6 distinction:

textScroll

AI output at U4 ≠ U6 field truth

Plain meaning:

AI can generate fluent claims, summaries, labels, classifications, explanations, safety judgments, and interpretations. But these are still U4 outputs. They become truth-valid only when source trace, context integrity, uncertainty, auditability, feedback, correction, and field validation are preserved.

Truth-discipline form:

textScroll

AI_truth_valid ⇔ source_trace + context_integrity + Au + FI + Ψ + Τ

Failure form:

textScroll

AI confidence↑ + source_trace↓ + U6 validation↓ ⇒ H_AI↑

Primary variables:

AI_claim, U4_output, U6_validation, source_trace, context_integrity, evidence_trace, claim_scope, uncertainty_trace, confidence_calibration, summary_integrity, field_effect_trace, correction_path, appeal_path, affected_node_feedback, H_AI, Φ_AI, Au, Au_eff, FI, BΣ, R, R_eff, L, Γ, Γ_AI, Π, Ξ, ℛ, Θ, Σ, Ψ, Τ

Diagnostic signature:

AI confidence rises while source trace, context integrity, uncertainty trace, field validation, and correction pathways decline. This indicates AI U4 truth drift.

Failure risk:

AI U4 claim drift, AI hallucination, AI summary distortion, AI context collapse, AI source erasure, AI synthetic certainty, AI classification overreach, AI field detachment, AI proxy truth substitution, AI narrative overcompression, AI confidence inflation, AI feedback suppression, AI legibility collapse, AI trust collapse, hidden debt accumulation.

Restoration priority:

Identify the AI claim, classify claim type and scope, restore source and evidence trace, restore context and uncertainty trace, validate against field effects where required, restore correction and appeal pathways, repair downstream debt, and validate reduced recurrence over time.

LAW-122 — AI Error Lag Law

Registry

LAW-124 — AI Rule-Stacking Law