HAiPECR — Human-AI Pre-deployment Evidence & Certification Record
HAiPECR — Human-AI Pre-deployment Evidence & Certification Record
HAiPECR — Human-AI Pre-deployment Evidence & Certification Record — 57% of in-scope requirements covered.
8 requirements · 7 in scope (1 enforced · 6 designed) · 1 out-of-scope (outside KYE™’s authority layer). The 57% is weighted over the in-scope base.
Source: Human-AI Institute / Markus Krebsz — HAiPECR (Human-AI Pre-deployment Evidence & Certification Record); listed on the OECD Catalogue of Tools & Metrics for Trustworthy AI (April 2023); aligned to UNECE ECE/TRADE/486 human-centric trade/standards guidance. · License: HAiPECR is an expert governance framework authored by Markus Krebsz / the Human-AI Institute. The KYE™ registry paraphrases each dimension's runtime-facing intent and cites the framework for mapping purposes only; it does not reproduce the framework text.
By category
| Category | Reqs | Enforced | Designed | Advisory | Deferred | Coverage |
|---|---|---|---|---|---|---|
| Dimension — accountability | 1 | 1 | 0 | 0 | 0 | 100% |
| Dimension — human oversight | 1 | 0 | 1 | 0 | 0 | 50% |
| Dimension — transparency & explainability | 1 | 0 | 1 | 0 | 0 | 50% |
| Dimension — privacy & data | 1 | 0 | 1 | 0 | 0 | 50% |
| Dimension — compliance & legal | 1 | 0 | 1 | 0 | 0 | 50% |
| Dimension — resilience & security | 1 | 0 | 1 | 0 | 0 | 50% |
| Deploy-gate verdict consumption | 1 | 0 | 1 | 0 | 0 | 50% |
| Dimension — ethics & fairness | 1 | 0 | 0 | 0 | 0 | 0% |
Every requirement → the KYE™ artefact that enforces it
| ID | Title | Status | KYE™ enforcement |
|---|---|---|---|
haipecr.accountability |
HAiPECR dimension — Accountability: a named, accountable owner is on record for the AI system before deployment | enforced | engines: internalaudit_events: kye.purpose.request.v1constitution_refs: constitution/00-INDEX.md |
haipecr.human-oversight |
HAiPECR dimension — Human oversight: a human can oversee, intervene in, and halt the AI system's actions | designed | engines: internal, internalaudit_events: kye.purpose.admissibility.v1constitution_refs: constitution/36-GOVERNEDUI.md |
haipecr.transparency-explainability |
HAiPECR dimension — Transparency & explainability: the reasons for a consequential AI decision are recorded and explainable | designed | engines: internal, internalaudit_events: kye.evidence.decision_map.v1constitution_refs: constitution/13-RESILIENCE-LOOP.md |
haipecr.privacy-data |
HAiPECR dimension — Privacy & data: personal-data use by the AI system has a lawful basis and is admissibility-checked at the moment of use | designed | engines: internalaudit_events: kye.purpose.admissibility.v1constitution_refs: constitution/31-DATA-GOVERNANCE-PACK.md, constitution/63-MEMORY-AUTHORITY-RAIL.md |
haipecr.compliance-legal |
HAiPECR dimension — Compliance & legal: applicable legal/regulatory conditions are bound to runtime rules the system must satisfy | designed | engines: internalaudit_events: kye.purpose.request.v1constitution_refs: constitution/70-FRAMEWORK-MAPPING-RAIL.md |
haipecr.resilience-security |
HAiPECR dimension — Resilience & security: the system behaves provably under degradation and has no single point of failure in its authority path | designed | engines: internalaudit_events: kye.evidence.pack.v1constitution_refs: constitution/13-RESILIENCE-LOOP.md, constitution/51-NO-SPOF.md |
haipecr.deploy-gate-verdict-consumption |
The HAiPECR deploy verdict is consumed as an action policy — a do-not-deploy verdict denies all consequential actions for that system | designed | engines: internal, internalaudit_events: kye.purpose.admissibility.v1constitution_refs: constitution/70-FRAMEWORK-MAPPING-RAIL.md, constitution/12-PURPOSE-PERMISSION.md |
haipecr.ethics-fairness |
HAiPECR dimension — Ethics & fairness: scoring the ethical acceptability and fairness of the AI model's behaviour | out-of-scope | (no enforcement cited) |