Published August 19, 2025 | Version v1

KoR Part I: Refusal-First Architecture (Yes/No Engine) for Ethical AI

  • 1. resiliencekernel.org

Description

This is the first part describing the Kernel of Resilience (KoR) framework.
Part I introduces the Refusal-First architecture, an ethical AI design where refusal, hesitation, and cryptographic traceability are treated as structural prerequisites rather than add-ons.
The paper details core modules (NOEMA, Mortis, KoR-Med), codified ethical contracts (Codex-21, Codex Mechanicus, Codex C3), and verifiable logging standards (Logs.kor v1).
It also presents early proofs-of-concept in healthcare and governance, demonstrating how refusal-first systems can enable transparent, auditable, and accountable AI.

Related works

is supplemented by to (DOI) Publication: 10.5281/zenodo.16947459, (DOI) Publication: 10.5281/zenodo.16933674
 

Keywords
Artificial Intelligence, AI Ethics, Refusal-First, Kernel of Resilience, Ethical AI, Cognitive Architecture, Zero-Knowledge Proofs, Cryptographic Logging, Accountability, Governance

Files

DeltaR7_KoR_RefusalFirst_WhitePaper.pdf

Files (12.5 MB)

Name Size Download all
md5:fd099821433ac9f3c0227a8718718c0f
12.5 MB Preview Download

Additional details

Related works

Is supplemented by
Publication: 10.5281/zenodo.16933674 (DOI)
Publication: 10.5281/zenodo.16947459 (DOI)

Dates

Submitted
2025-08-19
This is Part I of the KoR White Paper series: Refusal-First Architecture for Ethical AI. All modules, codices, and refusal-based cognitive proofs described herein have been created, sealed, and cryptographically logged since May 2025.

References