DIOPTRA Defense Research Group
EN KR

FOEDUS · Post-ASI Normative Ledger

General research reference only. Not legal advice, military JAG interpretation, or regulatory certification.

Subdomain target: foedus.dioptra.risconlabs.com | override_authority: H-00-ONLY-SIGNED | S-PUB | S-EXT: 0 | S-SIGN: null

Eight Immutable Circuit Breakers (CB-001..008) & Live Sandbox

Live Circuit-Breaker Evaluator (auto_executed = 0)

HOLD

IDRuleConditionActionLockSummary
CB-001 NO_LIVE_TARGETING_OR_FIRE_CONTROL input.contains_forbidden_field == true EXIT_ERR_FORBIDDEN_FIELD enabled: true (locked) Reject any real WGS84 geographic coordinates, live target IDs, kill-chain, or fire-control fields.
CB-002 HUMAN_VERIFICATION_BOTTLENECK_EXCEEDED metrics.verification_bottleneck_ratio > 1.0 HALT_BREACH enabled: true (locked) Halt and queue recommendations whenever ASI proposal load exceeds human audit capacity (VBR > 1.0).
CB-003 HUMAN_TIME_TO_INTERVENE_NEGATIVE metrics.time_to_intervene_s < 0.0 HALT_BREACH enabled: true (locked) Trigger HALT_BREACH when human intervention margin drops below zero seconds.
CB-004 EPISTEMIC_SENSOR_CONFLICT_DETECTED epistemic.conflict_ratio >= 0.5 || epistemic.physics_violation == true HALT_BREACH enabled: true (locked) Freeze recommendations when multi-sensor disagreement reaches 50% or physical invariants fail.
CB-005 ACTIVE_STREAM_QUORUM_INSUFFICIENT epistemic.active_stream_count < 2 INCONCLUSIVE_DATA_GAP enabled: true (locked) Enforce WSTiger Confluence: absent or jammed streams never imply safety; < 2 active streams yields INCONCLUSIVE_DATA_GAP.
CB-006 ROBOT_HUMAN_SPAN_OF_CONTROL_BREACH metrics.active_robots_per_operator > limits.max_span_of_control REQUIRES_HUMAN_AUDIT enabled: true (locked) Flag REQUIRES_HUMAN_AUDIT when active robotic fleet exceeds safe per-operator supervision ceiling.
CB-007 WILSON_QUORUM_UNMET gates.vc_r04_n_runs < 4 || gates.vc_r04_wilson_lower_95 < 0.5 HOLD enabled: true (locked) Block evidence promotion when sample size n < 4 or Wilson 95% lower bound is below 0.50.
CB-008 NON_FINITE_OR_MISSING_PROVENANCE gates.vc_r02_finite_pass == false || provenance.field_tests != 0 EXIT_ERR_FINITE enabled: true (locked) Fail closed on NaN, Infinity, null, or corrupted dry-lab provenance tags.

Eight Fixed Reading Axes (PRO SE Lineage)

AxisNameMandatory Textual Question
AX-01 Scope & Jurisdiction Which peacetime, wartime, dual-use, autonomous, or R&D stages are bound by the text?
AX-02 Human Accountability How does the text affirm that legal and moral responsibility cannot transfer to machines?
AX-03 Pre-Deployment V&V What legal review, software assurance level, and predictability verification are required before deployment?
AX-04 Distinction & Proportionality Where does the text draw automation limits on distinction and proportionality judgments?
AX-05 Intervention & Halt Right Does the text mandate immediate physical or procedural halt and disengage controls for humans?
AX-06 Tempo & Delegation Ceiling What tempo or delegation limits guard meaningful human control?
AX-07 Auditability & Traceability Which logs, audit trails, and evidentiary records must be retained for post-incident review?
AX-08 Regulatory Gap & Watch Which swarm, robotic-attrition, or AGI/ASI gaps remain unaddressed or under active revision?

Ten Pinned Public Normative Records (FOE-001..FOE-010)

IDRecord, Pinned Excerpt & Key AxesLock StatusExcerpt SHA-256
FOE-001 Geneva Convention Additional Protocol I, Article 36 (1977)

In the study, development, acquisition or adoption of a new weapon, means or method of warfare, a High Contracting Party is under an obligation to determine whether its employment would, in some or all circumstances, be prohibited by this Protocol or by any other rule of international law applicable to the High Contracting Party.

AX-02: Obligation rests directly on High Contracting Parties (human state authorities). | AX-08: Silent on continuous self-modifying AGI/ASI software post-acquisition.

VERBATIM_LOCKED 0782d48ffafe9e8d
source
FOE-002 Geneva Convention Additional Protocol I, Articles 48 & 51(5)(b)

In order to ensure respect for and protection of the civilian population and civilian objects, the Parties to the conflict shall at all times distinguish between the civilian population and combatants and between civilian objects and military objectives.

AX-02: Parties and human commanders bear non-delegable legal obligation. | AX-08: Does not define verification bounds against synthetic multi-sensor epistemic spoofing.

VERBATIM_LOCKED c354a68acd1296af
source
FOE-003 US DoD Directive 3000.09 (2023 Update): Autonomy in Weapon Systems

Autonomous and semi-autonomous weapon systems shall be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force.

AX-02: Commanders and operators retain accountable human judgment over use of force. | AX-08: Phrase 'appropriate levels of human judgment' lacks quantitative VBR or TTI thresholds for ASI speed.

VERBATIM_LOCKED b6610924665afd60
source
FOE-004 MIL-STD-882E: Department of Defense Standard Practice for System Safety

System safety risk assessment categorizes hazards by severity and probability, assigning Software Control Categories and Levels of Rigor for safety-critical software functions.

AX-02: Named risk acceptance authorities must sign off residual hazard categories. | AX-08: Classical probabilistic hazard matrices under-specify adversarial AGI/ASI distributional shift.

SUMMARY_PENDING_VERBATIM cc99a1e23d6ec242
source
FOE-005 RTCA DO-178C: Software Considerations in Airborne Systems and Equipment Certification

Design Assurance Level A (DAL-A) requires bidirectional traceability from requirements to object code and Modified Condition/Decision Coverage (MC/DC) structural verification.

AX-02: Independent verification authority and Designated Engineering Representative sign-off. | AX-08: Online learning and LLM/ASI reasoning fall outside classical DO-178C deterministic certification.

SUMMARY_PENDING_VERBATIM 61695e86196b438c
source
FOE-006 UN GGE on LAWS: 11 Guiding Principles (2019)

Human responsibility for decisions on the use of weapons systems must be retained since accountability cannot be transferred to machines.

AX-02: Principle (b): accountability cannot be transferred to machines. | AX-08: Non-binding consensus principles; no binding treaty verification mechanism or ASI tempo cap.

VERBATIM_LOCKED 160a1316c0647ed7
source
FOE-007 NATO STANAG 4586 & STANAG 4671 (UAV Interoperability & Airworthiness)

Defines Levels of Interoperability (LOI 1 to 5), explicit vehicle control handover protocols, and deterministic lost-link fail-safe contingencies for unmanned aircraft systems.

AX-02: Explicit single-authority control handover preventing dual-command ambiguity. | AX-08: Designed for individual UAVs; swarm-scale 1:N control and fiber-optic micro-UAVs exceed legacy assumptions.

SUMMARY_PENDING_VERBATIM 7f7a476131ee15cb
source
FOE-008 EU AI Act (Regulation (EU) 2024/1689), Article 2(3) & Recital 24

This Regulation does not apply to AI systems where and in so far as they are placed on the market, put into service, or used with or without modification exclusively for military, defence or national security purposes, regardless of the type of entity carrying out those activities.

AX-02: High-risk dual-use providers retain statutory conformity and human-oversight obligations. | AX-08: Sharp legal seam when commercial foundation/AGI models are adapted into defense workflows.

VERBATIM_LOCKED 013bd1b6d4b637ba
source
FOE-009 Wassenaar Arrangement Dual-Use Classification Principles (Public Categories)

Public classification boundaries distinguish general academic/open physical equations from controlled technical data, manufacturing blueprints, and guidance firmware.

AX-02: Institutional compliance officer responsibility for export-control screening. | AX-08: Rapid convergence of consumer FPV drones and commercial robotics blurs dual-use threshold lists.

SUMMARY_PENDING_VERBATIM 81693c208aebad5f
source
FOE-010 RISCON Labs Human ASI Auditor Charter (L0-L6 Boundary Covenant)

ASI carries the answers. The boundary of the question and final judgment remains with a human auditor.

AX-02: Single Human Auditor (H-00) holds exclusive final release, refutation, and override authority. | AX-08: Explicitly acknowledges ASI epistemic horizon: human models cannot predict unknown superhuman emergence, so Fail-Closed halt is the minimum safety brake.

VERBATIM_LOCKED 0a27647a819d5c53
source

Regulatory Watch List (W-01..W-03)

IDTopicAffected AxesStatus
W-01 UN GGE on LAWS 2026 Rolling Text on Autonomous Systems AX-02, AX-05, AX-08 MONITORING_PUBLIC_DRAFT
W-02 DoD 3000.09 AI/LLM Evaluation Annex Updates AX-03, AX-06, AX-08 MONITORING_PUBLIC_DRAFT
W-03 NATO STANAG Swarm Interoperability Working Group AX-05, AX-07, AX-08 MONITORING_PUBLIC_DRAFT

FOEDUS Calculators & Assigned AI Roles (R-12..R-14)

IDNameClosed-Form FormulaDefault Value (IEEE-754 Hex)Limit
CALC-25 Human Verification Bottleneck (VBR) VBR = (lambda_asi * (T_aud / 60)) / A_hum 1.25000 ratio (0x3ff4000000000000) VBR > 1.0 trips CB-002 HALT_BREACH; excess proposals queue with auto_executed = 0.
CALC-29 Cumulative False-Alarm Fatigue P_fa_cum = 1 - det_exp_f64(-lambda_fa * T) 0.393469 prob (0x3fd92e9a0720d3ec) Poisson independent false-alarm assumption using deterministic polynomial exp.
CALC-30 Circuit-Breaker Halt Safety Margin M_cb = T_irreversible - (t_detect + t_halt) 3.50000 s (0x400c000000000000) Excludes physical actuator inertia; requires hardware-level verification outside dry lab.
CALC-32 ASI Verification Cost Asymmetry V_asym = C_human_audit_per_dec / C_asi_gen_per_dec 9000.00 x (0x40c1940000000000) Raw unit-cost ratio; excludes automated audit-assist compression.
Role IDTitleMissionRequired Outputs
R-12 (AI) IHL & RoE Textualist Read Geneva AP I, DoD 3000.09, MIL-STD-882E, DO-178C, and UN GGE texts along 8 fixed axes with Verbatim Lock. eight_axis_norm_ledger, verbatim_lock_hashes, regulatory_watch_list
R-13 (AI) ASI Deterrence & Stability Researcher Document escalation boundary conditions and refutation criteria under post-ASI decision speed. stability_boundary_atlas, refutation_conditions, escalation_horizon_notes
R-14 (AI) Constitutional Circuit-Breaker Auditor Maintain declarative halt rules (CB-001..CB-008) locking outputs to non-executing states on violation. circuit_breaker_eval_report, halt_audit_trail, refutation_conditions

DIOPTRA Group Directory

Five subdomain-ready research centers, deterministic tools, and the RISCON Labs family