Prepared for BFH / Innosuisse project discussion

Deterministic verification infrastructure for AI governance.

DigiEmu Core & Proof provides reproducible decision states, execution receipts, replay verification and audit-ready evidence bundles for accountable AI systems.

Snapshot / Replay / Verify

1Define decision boundary
2Capture canonical snapshot
3Attach execution receipt
4Calculate state hash
5Replay verification
6Generate PASS / FAIL evidence
7Prepare audit-ready bundle

Open Standard. Commercial Infrastructure.

Transparent core standard. Professional enterprise adoption.

DigiEmu separates the public standard layer from commercial enterprise adoption. The Core specification is designed to be transparent, reviewable and interoperable. Baumgartner Digital Infrastructure provides the commercial infrastructure required for secure enterprise deployment, conformance testing, audit reporting, integration and support.

Proof concept

From AI decision to verifiable evidence.

Input

AI-assisted decision input

State

canonical_decision_state: reproducible

Hash

sha256: 8f4c...a91e

Verify

PASS

Input capturedBoundary definedCanonical state createdHash calculatedReplay verifier confirms integrity

BFH / Innosuisse project fit

From concept to validated applied research prototype.

Scientific contribution

Formalize deterministic knowledge-state boundaries, canonical snapshots and replay-verifiable transition chains for AI governance.

Applied prototype

Develop DigiEmu Proof as a working verification prototype that can generate PASS / FAIL evidence reports.

Use case validation

Evaluate the approach in regulated workflows such as medical AI, compliance, finance or public-sector AI governance.

Market relevance

Prepare DigiEmu for practical adoption as lightweight verification infrastructure for organizations using AI-assisted decisions.

Current validation tracks

Concrete partner-facing validation paths.

GazaCare AI

Medical AI traceability: decision fingerprint, QR/hash evidence and verifiable recommendation history.

Public Sector / CO₂

Decision-state reconstruction for public-sector workflows, climate reporting and accountable administrative AI.

TBN

Complementary trust architecture: TBN verifies agent trust; DigiEmu reconstructs decision state.

Antifragile.AI

Potential governance interoperability: DigiEmu as deterministic evidence layer for AI approval workflows.

Core 2.0 draft.4

Technical review candidate for BFH / partner feedback.

DigiEmu Core 2.0 draft.4 is positioned as a partner-testable milestone for reviewing deterministic snapshots, replay verification, schema boundaries, conformance reports and audit evidence bundles.

Request technical review package

Research problem

AI systems are documented, but their decision states are rarely reconstructable.

Regulated AI workflows need evidence that can be independently checked. Logs, screenshots and explanations are useful, but they do not automatically prove that a specific AI-assisted decision state can be reconstructed, replayed and verified.

Deterministic state

Define exactly what belongs inside the reproducible knowledge boundary.

Transition integrity

Verify that one state correctly follows another through a receipt-backed transition.

Audit evidence

Generate PASS/FAIL reports that make verification concrete and inspectable.

DigiEmu Core

A deterministic standard for AI decision-state boundaries.

Core describes how AI-related knowledge states are captured, canonicalized and separated from non-deterministic metadata. It defines what must be reproducible, what stays outside the hash and how AI governance evidence can be structured.

Canonical JSON snapshots
Inside-hash vs outside-hash boundary
SHA-256 state identity
Replay-ready decision states
EU AI Act traceability framing
DigiEmu Proof

A minimal verifier for deterministic execution integrity.

Proof checks whether snapshots, receipts and transition chains compose into a verifiable result. The output is intentionally concrete: PASS or FAIL.

Replay verification
Receipt continuity
Tamper detection
Negative tests for broken chains
Audit-ready evidence reports

Application fields

Designed for regulated workflows where decisions must remain verifiable.

Medical AI

Triage, risk assessment and human approval boundaries.

Legal & compliance

Reconstructable decision evidence for regulated workflows.

Finance

Auditable policy decisions and transaction-state verification.

Public-sector AI

Traceable AI systems for accountable administrative decision support.

Discuss a BFH / Innosuisse project setup.

DigiEmu can start as one applied research workflow: one decision boundary, one verification prototype, one evaluation report and one audit-ready evidence bundle.

contact@digiemu.ch