The Kretauer-Kelly-Bridge Model: Human-in-the-Loop Governance for Limiting Autonomous AI Systems under Uncertainty

This working paper introduces the Kretauer-Kelly-Bridge Model (KKB), a human-in-the-loop governance and decision architecture designed to limit autonomous AI system progression under structural uncertainty. The model addresses a core deficit of contemporary AI governance frameworks: the absence of an explicit ex-ante decision layer that determines whether system progression should occur at all. The KKB operates as a supervisory control and decision architecture preceding optimization, safety mechanisms, and alignment approaches. It systematically separates technical capability from normative legitimacy and treats progression as a consent-based decision rather than a default process. By formalizing responsible limitation and binding progression decisions to identifiable human responsibility, the Kretauer-Kelly-Bridge reframes AI governance from performance optimization toward accountability, legitimacy, and system maturity in autonomous socio-technical systems.