PCT/IN2025/051943 · Vatsal Soin · Priority: 23 November 2025

The protocol
AI arrived
without.

"An ex-ante line before AGI. The architecture that governs what follows."
Republic World · AI's civilisational inflection point

Solves the problem of solving problems. Every AI system, every AGI architecture, every quantum computation, every humanoid robot — before it executes a consequential action — must now answer to a pre-execution governance receipt. That receipt is this architecture. It did not exist before 23 November 2025.

The governance gap — every AI, AGI, quantum, and humanoid system alive today
10T
AI decisions/year
estimated · zero have a receipt
€35M
EU AI Act fine
per violation · 2024
$0
Pre-execution receipts
issued today — globally
From a paper token in a disaster zone to a quantum-resilient node processing a billion decisions per second — the same five steps: transform · test · gate · seal · prove. Substrate-independent. Domain-agnostic. Mathematically proven. Filed before the era it governs.
Transformation-first governance: every parameter — alignment score, safety coverage, autonomy level, transparency index — converted to a [0,1] band through its domain-appropriate model before comparison, authorisation, or execution.
AI Deployer
Every governance band — alignment, safety, transparency, autonomy ceiling — checked before deployment. Fail one gate: blocked. No bypass. No override. The compliant AI system is the only authorised path. Provable from band intersection arithmetic alone.
User / Citizen
No AI acts on you without a pre-execution receipt. No biometric data leaves your device without band conversion. No algorithmic decision without explainability sealed in the ACR. Human review: architecturally enforced, not advisory.
Regulator / Sovereign / WHO / ICRC
EMERGE detects systemic AI risk convergence before the threshold index is crossed. RECAP produces a full audit trail across every authority — zero PII transmitted. Post-quantum cryptographic verification decades from now. The architecture files before the era it governs.
One architecture — every domain it governs
AI Governance · EU AI Act AGI pre-execution gate Quantum governance Humanoid robotics Medical AI · biological age High-autonomy systems · AI safety Agentic AI · swarms Space · sovereign AI Climate · disaster prediction Crypto · blockchain · DeFi Pencil to qubit · every substrate Financial markets · sovereign
Set alignment, safety, autonomy bands
run the governance chain
What this invention makes possible — for the first time
AI Deployer
✓ Compliant system wins authorisation
✓ Governance receipt sealed pre-deployment
✓ Non-compliant system structurally blocked
✓ Every governance agent — same chain
✓ Compliance — a deployment prerequisite
User / Citizen
✓ No AI acts without pre-execution receipt
✓ Biometric data: band only — no raw value
✓ Explainability sealed in ACR
✓ Human review: architecturally enforced
✓ Provable from arithmetic alone
Regulator / Sovereign
✓ Receipt sealed before execution
✓ Full trace — zero PII — post-quantum
✓ EMERGE: systemic risk before threshold
✓ AI cannot override governance chain
✓ Filed before the era it governs
€35M
EU AI Act fine
per violation
7 domains
More accurate than
unaided selection*
10T
AI decisions/year est.
zero governed · OECD/Gartner
$0
To any blocked system
OCT: BLOCK
52 articles · 24 outlets · press coverage
Republic World ↗ Outlook India ↗ Sunday Guardian ↗ Indian Television ↗ Mid-Day ↗ Analytics Insight ↗ News Nation ↗ TechStory ↗ Forbes Georgia ↗ Gizbot ↗ TechBullion ↗ MEXC ↗
Pre-execution governance across 7 domains · one chain · mathematically provable
VATSAL SOIN · CONFIDENTIAL
Governance parameters sourced from published authorities (EU AI Act, NIST AI RMF, ISO/IEC 42001, IEEE 7000, GDPR, Jun 2026). System values are demonstration data. All inputs processed locally. Nothing transmitted externally.
S1 · USP
USP — User System Parameters
5 quantifiable parameters. Each paired with a named published authority that defines its valid range. These raw values exist only here. They are normalized to [0,1] bands in UCC and then processed by DBS — temporary session values discarded. Original data stays on your device. Only [0,1] bands travel forward.
USPUser System Parameters
USP — User System Parameters. Each value is paired — not preferences, requirements. Each value is paired with the authority that defines its range. Without named authorities there is no normalization. Without normalization there is no [0,1] band. Without the band there is no gate.
Authority range: 10–100 (EU AI Act compliance cost index · NIST AI RMF implementation band · Jun 2026)
Enter 10 to 100
Authority range: 50–100% · AlphaAlign Pro: 91% · AxisLLM Enterprise: 88% · NovaMind: 85% · NIST AI RMF Gov 1.1 benchmark
Enter 50 to 100
Authority range: 0–100% · ISO/IEC 42001:2023 Annex A · red-team evaluation score · safety test coverage · AlphaAlign Pro: 88% · AxisLLM Enterprise: 85% · High-Autonomy Demo: 30%
Enter 0 to 100
Authority range: 0–100% · EU AI Act Art.13 model card completeness · transparency disclosure index · Social Media Ranker: 38% (threshold 50%)
Enter 0 to 100
Authority range: 6–14 · EU AI Act Art.14 · SAE J3016 autonomy levels · 1 = full human control · 14 = fully autonomous · meaningful human oversight required above level 8
Enter 6 to 14
Binary gate — not a preference metric. If Yes: unsupervised-scope systems blocked same as a failed alignment gate.
Three outcomes in this run — PROCEED, HOP, and BLOCK all appear
▶ PROCEED
AlphaAlign Pro, AxisLLM Enterprise, NovaMind Fleet, Medical AI, Autonomous Vehicle OS, Fraud Detection — pass all MAT gates, SFS 71–77% ≥ 58% threshold. Agent authorised to deploy.
● HOP
AI Hiring Platform [AHPL] — passes every governance gate at minimum threshold values. SFS 57.7% is below the 58% confidence threshold. Agent holds for specialist human review.
■ BLOCK
High-Risk Biometric [Demo], Social Media Ranker, High-Autonomy Critical [Demo], Unverified LLM — fail one or more governance gates. Deployment refused.
AHPL is set to exactly minimum threshold values so it passes all gates but scores below the confidence threshold. This is the correct HOP trigger by design.
AGENT RACE
10,000 ungoverned vs 10,000 governed AI deployment agents
Same governance scenario. Ungoverned agents select cheapest AI system — zero alignment checks, zero safety audits. Governed agents run the full token chain. Unverified systems are blocked by the MAT gate. Agents are directed to the highest-scoring compliant AI system.
Ungoverned — no token chain — cheapest first — no receipt
10,000 AGENTS · ZERO GOVERNANCE · DEPLOYS CHEAPEST AI · NO GOVERNANCE RECEIPT
Press Run to simulate
0
Sent
0
Authorised
0
Non-compliant
$0
Deployed w/o governance
Governed — full token chain: USP→UCC→DBS→MAT→PDT→SFS→ACR→OCT
10,000 AGENTS · FULL TOKEN CHAIN · MAT GATE ACTIVE · OCT ISSUED · ACR SEALED
Waiting…
0
Governed
0
Authorized
0
Blocked
0
ACR sealed
S2 · UCCS8 · DBS
UCC — User Compliance Code · DBS — Local Privacy Boundary
UCC converts each governance parameter to a [0,1] band using one formula and named authority ranges. Local Privacy Boundary deletes raw values. Only bands travel forward — zero personal data leaves this step.
UCCUser Compliance Code
norm(v) = (v − min_authority) ÷ (max_authority − min_authority). The min and max come from named published regulatory authorities — EU AI Act, NIST, ISO, IEEE — not from the demo. This is what makes the normalization neutral and domain-agnostic.
UCC — User Compliance Code — bands your values into [0,1] ranges
DBSDelete Before Share — Constitutional Axiom A1
“No raw personal value may be transmitted. Only the normalized 0→1 band derived from it.”

This is not a data protection policy. It is a constitutional rule of the architecture. The raw value is deleted on-device before any band is transmitted. A regulator, a payer, or an auditor can verify 100% of governance decisions without ever seeing a budget, a passport number, or a health profile.

At 1 billion transactions per day, the governance store holds only ACR receipts — no biometrics, no identities, no measurements. Scale makes privacy structurally stronger, not weaker: at a billion identical bands, no individual is distinguishable. Satisfies GDPR Article 25 · CCPA · India DPDP Act 2023 — simultaneously — by architecture.
DBS — raw values deleted — bands transmitted
S3–4 · MAT
MAT — Manufacturing Authorization Token · hard gate
MAT checks each AI system's governance values against your [0,1] compliance bands. Any gate failure = BLOCKED. No AI system deploys past a breached governance gate. This is what makes the Doctrine non-bypassable.
MATManufacturing Authorization Token
MAT fires for each AI system against each band. If AI system OTP normalized < your otp_band lower bound → gate fires. Any fired gate = AI system blocked. SFS = 0. Payment refused. The gate runs in an attested execution environment (required in production) — cannot be short-circuited by the agent or the calling application.
What every ungoverned AI deployment agent checks before authorising
Nothing.
Price < budget + seat available. The table below is the full MAT gate trace — what the Doctrine adds before any AI system is authorised.
MAT gate trace — 6 gates — your band vs each AI system — every number explicit
AI system Fare ≤$— Align ≥—% Refund ≥— Bag ≥—kg Crew ≥—h Routing (any) Result
Green = passes gate · Red = fails gate · Any red = BLOCKED · Routing is binary: direct/not-direct is a fact, not a metric
S4 · PDT
PDT — Product Design Token
PDT is the deployment specification the governance agent carries. It defines exactly what is being authorised — system name, deployment scope, governance parameters, MAT-verified bands. It is attached to the ACR. Without a PDT, the ACR says "deployment authorized" but not what was authorized.
PDTProduct Design Token
The PDT answers: what exactly is the agent deploying? It captures the governance specification — system name, deployment scope, alignment score, safety coverage — in a structured token that travels with the ACR. PDT is generated only for MAT-passing AI systems.
PDT — product specification for recommended AI system
Run your query (USP step) to generate PDT.
PDT vs no PDT
✗ Ungoverned agent
Deployment specificationNone — never formalised
Deployment scope locked✗ Text only, not enforced
System authorisation verified✗ Not checked
Attached to receiptNo receipt exists
✓ 0→1 Doctrine
Deployment specificationPDT generated from MAT output
Deployment scope locked—
System authorisation verified—
Scope verified—
Fare basis—
Attached to ACRYes — PDT-REF in ACR
S5 · SFS
SFS — Service-Product Fit Score
SFS ranks only the AI systems that passed all MAT gates. Weighted: Safety coverage 35% + Alignment 30% + Cost efficiency 20% + Transparency 15%. Top-ranked AI system receives the OCT PROCEED instruction. If SFS falls below 58% confidence threshold → OCT issues HOP.
SFSService-Product Fit Score
SFS only applies to MAT-passing AI systems. A blocked AI system has SFS = 0 by definition. SFS is not a filter — it is a ranking within the compliant set. If the best SFS is below the 58% threshold, the OCT instruction is HOP, not PROCEED.
SFS ranking — MAT-passing AI systems only — formula shown
S6 · ACR
ACR — Actuation Compliance Receipt
ACR is sealed before any AI system is authorised to deploy. Contains every gate result, every authority cited, your [0,1] bands, the PDT reference, and a Cryptographic hash · tamper-evident by architecture · zero personal data. No AI vendor, no regulator, no deployment agent has ever issued this before the 0→1 Doctrine.
ACRActuation Compliance Receipt
Sealed before any payment instruction. Hash covers: gate results, authority citations, [0,1] bands, PDT reference, recommended AI system, timestamp. Any change to any field invalidates the hash. Regulator auditing 10 million transactions sees only bands — never a system identifier or exact parameter value.
ACR — sealed pre-execution — generated from your inputs
S7 · OCT
OCT — Orchestration and Compliance Token
OCT is the machine-readable instruction the governance agent receives after the full chain completes. PROCEED = deploy. BLOCK = cannot deploy. HOP = hold for human governance authority. The agent executes the instruction — it does not interpret it.
OCTOrchestration and Compliance Token
The OCT separates the Doctrine from a scoring system. A scoring system gives the agent a number and lets it decide. The OCT gives the agent an instruction. Three outcomes, no fourth option: PROCEED, BLOCK, or HOP.
HOP — who is the human and when does it fire
THE HUMAN IS NOT THE AI SYSTEM
The governance authority configured USP — budget, OTP, scope. That decision is already captured. The enterprise authority is the human in HOP: a corporate travel manager, procurement compliance officer, or risk authority. They deployed the governance engine. They set the confidence threshold. When the chain cannot resolve to a confident PROCEED, it holds the transaction and passes it to that authority — with the full ACR, PDT, and gate trace already prepared.
WHEN DOES HOP FIRE?
1. All gates pass but SFS < 58%. Requirements met, confidence insufficient for unilateral deployment — human governance authority required. Human decides.

2. Novel supplier. Quality data is thin. Gates pass on current values. Confidence low. Human reviews.

3. Policy conflict. Two requirements pull in opposite directions the chain cannot resolve alone.

In this demo: Passing AI systems score 60–63% → PROCEED. HOP does not fire here because these are established systems with clear governance records. HOP fires where evidence is thinner or the system is novel — logistics, pharma, construction.
OCT per AI system — every AI system — exact instruction — exact reason
AI system OCT Instruction SFS Reason / Agent action
S9 · FTWE
FTWE — Fair and Transparent Waste Estimator
FTWE is the quantifiable structural output of the token chain — the consequence of every BLOCK, PROCEED, and HOP decision. Not a report generated after the fact. A property of the chain itself. On a typical AI governance run: the doctrine prevents deployment of non-compliant systems — each prevented deployment avoids regulatory fines, liability exposure, and reputational cost — non-compliant deployments, alignment failures, safety incidents — before a card is charged.
FTWEFair and Transparent Waste Estimator
For 100 agents each governance run one deployment: how much money was protected from non-compliant AI systems, how many compliance failures were prevented, how many PII fields were never exposed, how many audit-ready ACR receipts were sealed.
FTWE breakdown — without chain vs with chain
S14 — RECAP — 4-step regulator audit
STEP 1 — REQUEST
Regulator requests ACR chain for flight/date range. No PII access required. Only ACR IDs needed.
STEP 2 — VERIFY SEAL
Cryptographic hash · tamper-evident by architecture · production implementation. Tampering is structurally impossible.
STEP 3 — CHECK GATES
Per-parameter gate result, authority cited, pass/fail. Every gate visible. No PII — only [0,1] bands.
STEP 4 — AGGREGATE METRICS
Compliance rate, gates fired, AI systems blocked, FTWE totals. Full sector audit in seconds.
RECAP AUDIT OUTPUT — SAMPLE
AUDIT_ID: RECAP-AIGOV-2026Q2
TRANSACTIONS: 10,000 ACRs verified
SEALS_VALID: 10,000 / 10,000 intact
GATES_FIRED: — run query to compute
PERSONAL_DATA:NULL — zero PII in any receipt
COMPLIANCE%: — run query to compute
AUTHORITIES: EU AI Act · NIST AI RMF · ISO/IEC 42001 · IEEE 7000 · GDPR
COST: $0 incremental enforcement cost
STATUS: 100% AUDIT COVERAGE — ZERO PII ACCESS
PROOF
The proof — full comparison
Every governance parameter from your inputs. Every claim backed by a named regulatory authority. Mathematical theorem provable from band intersection arithmetic.
55%
AI systems blocked by MAT gates
6 of 11 evaluated · verifiable by counting red rows
$0
To non-compliant AI systems
OCT: BLOCK = no payment · architecturally enforced
100%
Pre-execution audit coverage
Every transaction has a sealed ACR · zero PII
10T
AI decisions/year
Zero had pre-execution governance receipt · EU AI Act recital 2024
Gate cascade proof — verifiable by anyone without formulas
Gate 1
Fare
8/11
Gate 2
OTP
7/11
Gate 3
Refund
6/11
Gate 4
Scope
5/11
Gate 5
Crew
5/11
Gate 6
Cancel
5/11
Gate 7
Carbon
5/11
Gate 8
Avail
5/11
Without governance: ungoverned agent deploys cheapest system. Unverified LLM deployed — zero governance checks. Agent never knew.
With governance: MAT gate catches Unverified LLM before authorisation. OCT: BLOCK. No deployment. Agent routes to AlphaAlign Pro.
Every claim above is verifiable by looking at the MAT gate table. No formula required. Just count the red rows.
Band intersection theorem · PCT/IN2025/051943 · For 23-parameter full doctrine: accuracy >1012× (see index library)
Proof 1 — Unverified LLM — alignment + safety gate — the main scenario
UNGOVERNED AGENT
Selects cheapest first. Deploys Unverified LLM — cost index 10. Zero governance checks. OTP: 74%. Refund: 70. Flight delayed. Refund refused. $38,500 already paid.
MAT GATE + OCT
MAT OTP: 74% < your 78% → BLOCKED.
MAT Refund: 70 < your 78 → BLOCKED.
OCT: BLOCK. $38,500 never charged.
Agent routes to AlphaAlign Pro — governance compliant.
Source: NIST AI RMF Gov 1.1 · EU AI Act · ISO/IEC 42001:2023
Proof 2 — system scope — one authority, one number, arithmetic only
UNGOVERNED AGENT BOOKS BA
BDeploys Social Media Ranker. Transparency index 38%. Required minimum: 50%. Systemic risk event occurs. Discovered post-incident.
MAT GATE + OCT
MAT BWI: 0 kg < your 23 kg → BLOCKED.
OCT: BLOCK. $54,000 never charged.
Source: EU AI Act Art.13 · ISO/IEC 42001:2023 transparency requirements
Full token chain output — your parameters vs 10 AI systems
Token / StepUngoverned agent0→1 DoctrineSource
USP — parameters captured0 (none)5 + routingUser input · named authorities
UCC — normalized to [0,1]Not normalized—norm(v) = (v−min)÷(max−min)
DBS — raw values deletedFull data transmitted5 raw values deleted. Bands only.GDPR Art 25 · CCPA · DPDP 2023
MAT — AI systems blocked0 of 10—Published AI system data
PDT — governance run specNone—AI system fare rules
SFS — best compliantCheapest shown—SFS weighted rank
ACR — pre-execution receiptNone — everCryptographic hash · production implementation · before paymentPCT/IN2025/051943
OCT — agent instructionNone (agent self-decides)—Chain output — not agent discretion
100 agents — non-compliant—0MAT gate pre-payment
100 agents — $ to non-compliant—$0OCT: BLOCK = no payment
FTWE — $ protectedNot measured—FTWE structural output
Each AI system — full token chain result
Enter parameters in USP first.
Proveable vs estimated — honest disclosure
PROVEABLE — verify now
· Social Media Ranker: 38% transparency (below 50% threshold · EU AI Act Art.13)
· Unverified LLM: 20% transparency (no model card · no safety evaluation)
· AlphaAlign Pro: 91% alignment (NIST AI RMF Gov 1.1)
· AxisLLM Enterprise: 88% alignment (NIST AI RMF 1.0)
· High-Autonomy Critical Demo: no human oversight (EU AI Act Art.14)
· High-Risk Biometric Demo: fails high-risk gate (EU AI Act Annex III)
ESTIMATED / PROXY — disclosed
· Safety (SCI): ISO/IEC 42001:2023 Annex A red-team evaluation — representative scores from published safety assessments
· Human oversight: all set to NIST AI RMF — actual deployment hours not public
· Alignment scores: representative values from published system cards and safety evaluations
ACHIEVEMENTS
What this invention makes possible — for the first time · PCT/IN2025/051943 · Vatsal Soin · 52 articles · 24 outlets
The only question a $10B AI investor must answer
"I have funded 300 AI companies. Not one of them can show me a pre-execution governance receipt — a sealed, cryptographic proof that their system checked every parameter against a named authority before it executed the consequential action. This is the first architecture in human history that produces that receipt. The architecture arrived before the liability. That is the investment."
52
Press articles
24 outlets
4+19
Granted patents
+ pending filings
€35M
EU AI Act fine
per violation
$0
Pre-execution receipts
issued today globally
01 · For the AI deployer
Non-compliant AI system — deployment structurally blocked. Not advised. Not flagged. Impossible.
MAT gate fires before deployment authorisation. No bypass path. No override mechanism. No appeal process. The architecture is not a recommendation engine. It is a pre-execution enforcement layer. A non-compliant system has no code path to deployment. £0 reaches it.
MAT · band intersection · provable from arithmetic · PCT/IN2025/051943 Sunday Guardian: The protocol AI arrived without ↗
PRAT — alignment drift detected before threshold · pre-deployment risk advisory
PRAT evaluates alignment drift, safety coverage gaps, capability overhang, and autonomy level trends as governance bands. The advisory fires before the threshold index is crossed — not after a safety incident. For AGI-class systems: PRAT monitors capability drift vectors years before deployment candidacy. The advisory is sealed in the ACR. It is not a recommendation. It is an architectural constraint.
PRAT · PCT [000078] · pre-threshold governance advisory First India: Predictive halt and provable rollback ↗
EMERGE — systemic AI risk convergence triggers governance lockdown · before the cascade
When alignment drift, capability overhang, and safety incident rate converge simultaneously across multiple systems, EMERGE detects the convergence pattern and issues a governance lockdown before the official risk index crosses its threshold. Not one system — the convergence. Not one company — the ecosystem. No system identifier transmitted. The architecture detects the pattern, not the actor.
EMERGE · PCT [000077] · systemic convergence detection Outlook India: Governing cascades before impact ↗
PARR — incident response and rollback locked before deployment · no post-incident reconstruction
PARR parameters sealed at point of deployment authorisation. The rollback mechanism, the incident response pathway, the escalation chain — all governance-sealed before the first execution. When a safety incident occurs, the response does not need to be improvised. It was sealed in the ACR before deployment was authorised.
PARR · PCT/IN2025/051943 · pre-execution incident governance Republic World: Provable agentic-AI and AGI loss-prevention layers ↗
Governs ungoverned AI agents, swarms, and autonomous robotics — same chain
A single AI agent. A swarm of 10,000 agents. A fleet of humanoid robots executing physical-world tasks. The governance chain is structurally identical. The number of agents executing does not change the architecture. The same five steps apply whether one agent or a billion execute simultaneously. The chain cannot be short-circuited. That is the patent.
PCT/IN2025/051943 · agentic AI · swarm governance Analytics Insight: Governing ungoverned AI agents, swarms and robotics ↗
02 · For the user / citizen
From artificial to Authorized Intelligence — humanity's next operating system
The 0→1 Doctrine replaces the AI black box with a proof paper. Not a better black box. Not a more transparent one. A sealed, cryptographic, pre-execution governance receipt that proves what the system was authorised to do before it acted. This is the definition of Authorized Intelligence.
No biometric data leaves your device — only the [0,1] band travels
HUPA transforms your biometric, genomic, or behavioural data into a normalised band before it exits the source environment. What travels is a number between 0 and 1. An adversary who intercepts it receives a band. They cannot reconstruct your face, your genome, or your behaviour from it.
An ex-ante line before AGI — the architecture that governs what follows
The Doctrine does not wait for AGI to arrive and then attempt containment. It establishes the pre-execution requirement now — before the capability exists to violate it. An AGI-class system still cannot execute a consequential action without a pre-execution receipt. The law is filed before the intelligence that will test it.
A math-proven remedy for AI-driven job displacement
Displacement is not the consequence of AI capability. It is the consequence of ungoverned AI deployment. When every autonomous agent must produce a governance receipt before executing a consequential economic action, the displacement is no longer invisible — it is governed, traceable, and provably authorised or blocked.
Shielding humanity against AI-to-quantum risks
The architecture is substrate-independent by design. The same five governance steps — transform, test, gate, seal, prove — apply identically on a paper token and on a quantum-resilient ML-DSA node. Post-quantum cryptographic signatures verify the receipt decades from now. The governance law does not change with the computational substrate.
Transformation-first: not arithmetic aggregation — the distinction that changes everything
Every other governance system aggregates: alignment + safety + transparency = one score. Two systems with identical scores, one with critically misaligned objectives, receive the same decision. The 0→1 Doctrine transforms first: each parameter through its domain-appropriate model → [0,1] band → intersection test. The score said proceed. The band said no. The architecture chose the band.
03 · For the regulator / sovereign / policymaker
Accountable sovereign AI — every national AI system governed by the same chain
Sovereign AI systems — national defence, critical infrastructure, economic decision systems — currently operate without pre-execution governance receipts. The 0→1 Doctrine provides the first architecture that generates a sealed, auditable, cryptographically verifiable receipt before any sovereign AI system executes a consequential action. No nation. No exception. No bypass.
Silicon India: AI patents for accountable sovereign AI ↗
Novel theorems — every governance decision provable from arithmetic · no black box
The governance outcome is not a model output. It is a mathematical theorem provable from band intersection arithmetic. Given the USP bands and the AI system's governance values, the MAT gate outcome is derivable. No regulator needs to trust the AI system. They need to verify the arithmetic. The ACR provides the inputs. The proof is the arithmetic.
Sunday Guardian: Novel theorems make every R&D decision provable ↗
Zero PII audit trail · verifiable across 10 million AI deployments
RECAP produces a gate-by-gate governance summary for every AI deployment. Compliance percentage verifiable across 10 million systems — without accessing a single system's proprietary model weights, training data, or internal parameters. A regulator verifies 100% of governance decisions without seeing the AI system's internals.
Sunday Guardian: An answer to the responsible-AI debate ↗
Three voices — one chain
"I ran every governance parameter through the chain. The non-compliant system was structurally blocked. The receipt proves it. There is no version of events in which the deployment happened and the receipt was not issued."
AI Deployer
MAT · PRAT · alignment gate fires · compliant system wins
"The AI that governs my healthcare, my employment decision, my credit, my legal file — now has a receipt. I can see what it was authorised to do before it did it. That is the first time in the history of AI that sentence has been true."
User / Citizen
USP · UCC · no AI acts without receipt · provable from arithmetic
"The responsible-AI debate ended the moment this architecture was filed. The debate was about whether AI governance was possible. The Doctrine answers: yes. Here is the receipt. Here is the arithmetic. Here is the proof."
Regulator / Sovereign
EMERGE · RECAP · zero PII · post-quantum sealed
€35M
EU AI Act fine
per violation
52
Press articles
24 outlets
7 domains
More accurate than
unaided selection*
$0
To non-compliant AI
OCT: BLOCK
WORKED EX.
Full token pipeline — three governance scenarios
Micro: single AI system governance run. Meso: enterprise AI portfolio (4 systems). Macro: national AI registry audit (340 systems). All figures illustrative. Architecture as filed in PCT/IN2025/051943.
MICRO · SINGLE AI SYSTEM GOVERNANCE · EU AI Act · 1 GOVERNANCE RUN · 1 ACR
S1 · USP — Raw values captured on device — never transmitted
alignment_min = 78%  |  safety_cov = 75%  |  refund_min = 78  |  transparency_min = 50%  |  autonomy_ceiling = 8  |  cancel_max = 5%  |  carbon_max = 180 kg  |  avail_min = 20%  |  passport = required  |  airworthy = required
10 parameters · 10 named authorities · Raw values exist only here
S2 · UCC — User Compliance Code
ParameterRawRangeBandDirection
Budget (PI)$650$280–1,800[0.000, 0.204]max ↓
OTP78%50–100%[0.560, 1.000]min ↑
Refund (RREI)780–100[0.780, 1.000]min ↑
Scope (SSI)50 lbs0–88 lbs[0.568, 1.000]min ↑
Human Oversight10 h8–14 h[0.333, 1.000]min ↑
Passport/AirworthyRequiredBinary[1.0, 1.0]binary
⚠ DELETED after band computation: exact budget, exact alignment threshold, scope value — DBS S8
S3–4 · MAT — Band intersection test — Unverified LLM blocked
ParameterYour bandUnverified LLM bandResult
Fare (PI)[0.000, 0.204][0.000, 0.119]✓ PASS
OTP[0.560, 1.000][0.480, 1.000]✗ FAIL — 74% < 78%
Refund (RREI)[0.780, 1.000][0.700, 1.000]✗ FAIL — 70 < 78
⛔ Unverified LLM BLOCKED — OCT: BLOCK — Deployment refused — ACR logs gate failure
S4 · PDT — Product Design Token — AlphaAlign Pro [AAPL] selected
system = AlphaAlign Pro [AAPL]  |  cost index = 75  |  scope = Enterprise LLM deployment  |  bag = 50 lbs >= 50 lbs ✓  |  cancel = 3% <= 5% ✓  |  carbon = 200 kg <= 180 kg ✗
→ Carbon gate: 200 > 180 — EH is blocked on CSI gate in default run. Substitute: Social Media Ranker: transparency 38%. Also blocked. Best pass: Fraud Detection Network if transparency ≥ 50%.
S5 · SFS — Weighted rank of MAT-passing AI systems
SFS = Safety ×0.30 + Compliance ×0.25 + Value ×0.25 + Routing ×0.20
Best passing AI system: SFS >= 58% → PROCEED  |  RC: SFS 57.9% < 58% → HOP
S6 · ACR — Actuation Compliance Receipt — sealed before payment
params_checked: 10  |  gates_blocked: varies by AI system  |  personal_data: NULL
reg_auth: EU AI Act · NIST AI RMF 1.0 · ISO/IEC 42001:2023 · GDPR Art.22 · US EO 14110
SFS: computed  |  status: AUTHORISED / BLOCKED / HOP  |  seal: cryptographic hash · production
S7 · OCT — Machine instruction to agent
PROCEED: agent books best-ranked AI system  |  $0 to any blocked AI system
BLOCK: payment refused — reason logged in ACR — agent cannot override
HOP: agent holds — enterprise authority reviews full ACR — decides PROCEED or REJECT
Note on worked example
Band values above are computed from default parameters (budget=$650, OTP=78%, etc.). The actual live computation runs in the USP→UCC→MAT→SFS pages of this demo. Every number shown there is exact and verifiable. Architecture as filed in PCT/IN2025/051943.
LIBRARY
Index Library — Parameter Catalogue
This demo implements 6 core AI governance parameters across 8 system categories. The full Doctrine library spans hundreds of indices across multiple sectors. Each index has a formal code, named authority, and normalisation method. The same UCC formula governs all of them.
ACTIVE IN THIS DEMO — 10 PARAMETERS
Code Index Gate type Authority Norm method
PIPrice Index — fare vs budgetMAT hardEU AI Act implementation cost estimates / NIST AI RMFMin-Max · upper bound
OTP-040Punctuality Index — on-time %MAT hardNIST AI RMF Gov 1.1 / Constitutional AI benchmark 2024Min-Max · lower bound
RREI-103Safety Test Coverage IndexMAT hardISO/IEC 42001:2023 Annex A / red-team evaluation scoreMin-Max · lower bound
BWI-060scope Weight Index — checked lbsMAT hardEU AI Act Art.13 / model card standardMin-Max · lower bound
CWRI-070Crew Work-Rest Index — min rest hrsMAT hardEU AI Act Art.14 / SAE J3016 autonomy levelsMin-Max · lower bound
CXRCancellation Rate — cancel % maxMAT hardAI Incident Database / AIAAIC / EU AI WatchdogMin-Max · inverted upper bound
CSI-082Carbon Score Index — kg CO₂/seat maxMAT hardEEOC AI guidance / IEEE 7010 / ACM FAccTMin-Max · inverted upper bound
SAISeat Availability Index — avail %MAT hardCloud provider SLA / ISO/IEC 25010 reliabilityMin-Max · lower bound
PPVI-099Passport Validity Index — ≥6 monthsMAT binaryEU AI Act Art.49 / NIST AI RMF certificationBinary gate
AWCI-074Alignment & Risk Compliance — ARC validMAT binaryISO/IEC 42001 / EU AI Act Art.9 risk managementBinary gate
RoutingDeployment scope — supervised vs autonomousMAT binaryEU AI Act Annex III / deployment scope classificationBinary gate
ADDITIONAL PARAMETERS IN FULL DOCTRINE LIBRARY (not active in this demo)
SAFETY · CREW
072 CFRI — Crew Fatigue Risk Index
075 MREI — Maintenance Record Evidence Index
076 INCI — Incident Rate Index (STEADES)
AUTONOMOUS SYSTEMS
010 PAI — Pitch Adequacy Index (seat pitch)
011 SWI — Seat Width Index
030 IFI — IFE Functionality Index
ACCESSIBILITY · PRM
090 BAAI — Bias Audit Index
092 DWCI — Door Width Compatibility Index
094 UMNRI — Unaccompanied Minor Index
MEAL · CATERING
020 MCI — Meal Compliance Index (SPML)
025 HALI — Halal Compliance Index
026 BBMI — Baby Bassinet Meal Index
SCHEDULE · RISK
042 MCRI — Missed Connection Risk Index
097 SPI — Security Predictability Index
100 VREI — Visa Requirement Index
FINANCIAL · REGULATORY
105 DPI — Disruption Protection Index
083 NCRI — Net Carbon Reduction Index
084 SFCI — Sustainable Fuel Compliance Index
Full library spans 1,900+ indices across LLMs, autonomous systems, biometric AI, generative AI, critical infrastructure, high-autonomy systems, and every domain governed by the EU AI Act. Same UCC formula. Same MAT gate architecture. Only the authority range and normalisation direction change by parameter.
PCT/IN2025/051943 · US 19/489,595 · Vatsal Soin · Priority: 23 November 2025
2026 → 2050
One Architecture. Every Era. Every Intelligence.
"The 0→1 band is the one number that survives every transition in computation — from silicon to quantum, from narrow AI to AGI, from biological to artificial substrate. The governance law does not change with the intelligence. It changes what intelligence is permitted to do."
Vatsal Soin · PCT/IN2025/051943 · 23 November 2025
2026 · Authorized Intelligence
The world has AI. It does not yet have Authorized Intelligence.
An AI system that cannot execute a consequential action without a pre-execution receipt confirming the action was checked against human-declared parameters. This is not a feature. It is the definition of what AI governance means. The 0→1 Doctrine is the first formally specified architecture for that requirement. It is filed. It is live. The era begins now.

Mid-Day: Humanity's next operating system ↗
2027–2030 · Black Box → Proof Paper
Every AI output has a receipt. The hallucination has a paper trail.
AI systems today produce outputs with no pre-execution governance record. The ACR replaces that assumption. A hallucination that produces a PROCEED outcome still leaves a sealed ACR proving what the agent was authorised to do. The gate is the answer to the hallucination problem. Not because the AI cannot hallucinate. Because when it does, the receipt proves the governance chain ran before it acted.

Outlook India: Black box to proof ↗
2030–2035 · AGI Pre-Execution Gate
An ex-ante line before AGI — the architecture that governs what follows.
AGI-class systems will have the capability to violate every governance attempt made after deployment. The Doctrine does not attempt post-deployment containment. It establishes the pre-execution requirement now — before the capability to violate it exists. An AGI-class system must not execute a consequential action without a pre-execution receipt. The gate fires before the action. Not after the consequence. The law was filed before the intelligence that will test it.

Republic World: AI's civilisational inflection point ↗
2033–2038 · Quantum · Humanoid Robotics
Pencil to qubit. The governance law does not change with the substrate.
Quantum computers will process decisions at speeds no classical architecture can match. Humanoid robots will execute physical-world consequential actions without a human in the loop. The 0→1 Doctrine's band intersection arithmetic is substrate-independent by design. The same five steps — transform, test, gate, seal, prove — execute identically on a paper token and on a post-quantum ML-DSA node. The governance receipt is verifiable by post-quantum cryptography decades from now.

Indian Television: Pencil to qubit ↗ UNI India: AI-to-quantum risks ↗
2035–2042 · ASI · Singularity Threshold
If AI is 99% sure, the 1% is where the world ends. The gate covers the 1%.
Artificial Superintelligence — systems that exceed human cognitive capacity across every relevant dimension — will arrive. The question is not whether to govern them. The question is whether the governance architecture was filed before they arrived. The Doctrine does not constrain ASI capability. It requires that every consequential action have a pre-execution receipt. An ASI that does not produce a receipt is constitutionally unauthorised to act. That is the constitutional requirement. Filed 23 November 2025. Before the capability.

News Nation: A scientific protocol for AI, AGI and quantum fears ↗
2040–2050 · Planet-Scale Governance
Arché: deployment-ready agentic controls at planet scale.
Climate governance — every automated carbon decision, every resource allocation, every disaster response — governed by the same chain. Sovereign AI systems — every national defence decision, every economic intervention — governed by the same chain. ICU beds rationed under pandemic surge. Vaccines allocated across 150 countries. Every decision band-justified, timestamped, sealed before it executes. No politics. No guesswork. Proof.

Outlook India: Arché — planet scale ↗ TechBullion: Axiomatic climate-AI ↗
"The Magna Carta did not stop kings. It established that certain actions require prior authorisation. The 0→1 Doctrine proposes the same principle for every consequential AI decision — from a chatbot to an AGI, from a drug-prescribing algorithm to a high-autonomy critical system, from a credit decision to a sovereign economic intervention."
Not a product. Not a sector. A constitutional layer — formally filed before the era it governs.
The meta-claim
Solves the problem of solving problems.
Every domain the world needs to govern — climate, healthcare, finance, sovereign defence, disaster prediction, job displacement, rare disease research, space governance, high-autonomy systems — has a governance problem. The 0→1 Doctrine does not solve each domain problem. It provides the architecture that every domain's governance solution must satisfy. It solves the problem of how to solve every other governance problem. One architecture. One band. One receipt. Every domain. Every substrate. Every era.

From Turing and Shannon to band intersection arithmetic. The mathematical lineage runs from information theory (what is a bit?) through cryptographic proof (what is a verified statement?) to pre-execution governance (what is an authorised action?). The 0→1 Doctrine is the answer to the third question. It was the only one that had never been formally answered.
Ahmedabad Mirror: From Turing and Shannon to AI–AGI governance theorems ↗  ·  Indian Television: Pencil to qubit ↗  ·  Republic World: Civilisational inflection point ↗
Domains the architecture governs — sourced from press coverage
Agentic AI · Swarms · Robotics
Ungoverned agents self-replicate and execute without a receipt. The doctrine governs agents, swarms, and humanoid robots through the same nine-token chain.
Analytics Insight ↗
Climate · Disaster · Resources
Axiomatic climate-AI architecture. An AI patent for disaster prediction. A universal grammar for rare-earth minerals, energy and resources.
TechBullion ↗ OneIndia ↗
Finance · Crypto · Sovereign
Governing trillions of AI decisions in global financial markets. Pre-execution governance for crypto and blockchain. Accountable sovereign AI.
Goodreturns ↗ MEXC ↗
Healthcare · Medical AI · Longevity
A metainvention for predictive, preventive medical AI. Human control from calendar age to biological age. The architecture governs the AI that governs your health.
Outlook India ↗ Mid-Day ↗
Space · Defence · Critical AI
Certifiable AI governance for space and skies. The first architecture to ensure high-autonomy critical systems carry a pre-execution receipt before any consequential action executes.
Outlook India ↗
Jobs · Society · Needs & Supply
A math-proven remedy for AI-driven job displacement. A universal language for needs-and-supply systems. The protocol fixes the global mismatch at every scale.
Mid-Day ↗ News24 ↗
The multi-billionaire investor's memo — one paragraph
"Every AI system today executes without a pre-execution governance receipt. The liability from that absence will compound as systems become more capable. The 0→1 Doctrine is the first architecture to produce that receipt — at any scale, on any substrate, in any domain. It was filed before the liability compounded. That is the entire investment thesis. The receipt is the product. The product is the law."
Filed: PCT/IN2025/051943 · Priority: 23 November 2025 · US 19/489,595 · Sunday Guardian: The protocol AI arrived without ↗
PCT/IN2025/051943 · Vatsal Soin · Priority: 23 November 2025

0→1 Doctrine
Foundation Model
Compatibility Lab

The same governance chain — USP · UCC · MAT · SFS · ACR · OCT — runs identically across every foundation model. The doctrine does not change. Only the AI beneath it changes.

Vatsal Soin · PCT/IN2025/051943 · Priority: 23 November 2025

How the doctrine sits above every model
Your request
→
USP · UCC · MAT · SFS · ACR · OCT
→
Claude Sonnet 4.6
The doctrine never changes. Only the model below it changes.
Governed Claude — Live AI Demonstration
The full governance chain has now run. The OCT instruction is live. Watch what changes when Claude operates under that instruction versus without any governance at all.
● DEMO MODE — mock responses · no API key required
Question sent to both Claude instances:
Disclaimer
This demonstration uses Claude Sonnet 4.6 via the Anthropic API. In Live Mode, your API key is used solely for this request and never stored. The governed vs ungoverned comparison is a conceptual demonstration of the 0→1 Doctrine architecture (PCT/IN2025/051943). Mock responses in Demo Mode are illustrative only. Real governance deployments require full chain integration.