Event-Anchored Trust in AI-Powered Assessment: Why Trust Collapses Faster than It Repairs

Trust in AI-powered grading systems can shatter overnight after a single disputed score, yet rebuilding that trust takes months. This paper develops a framework to explain this “fast-collapse, slow-repair” asymmetry. We introduce Event-Anchored Trust (EAT), reconceptualizing trust as an event-contingent judgment rather than a stable attitude, and the Coupling-Threshold-Hysteresis (CTH) dynamical systems framework, which reveals how perceived fairness, perceived control, and psychological safety interact as a self-reinforcing system that can abruptly flip from trust to distrust when a critical threshold is crossed. The framework demonstrates a hysteresis effect: returning to trust demands greater effort than the initial collapse. We propose an “Explain–Re-evaluate–Respond” synergistic repair framework, showing that at least two coordinated strategies are necessary and that contradictory repair signals worsen the lock-in. The framework directly informs institutional crisis protocols and pre-deployment safeguards, providing administrators with a falsifiable logic for managing AI assessment trust crises. This article is a conceptual paper employing structured theory synthesis.