publications

publications by categories in reversed chronological order. generated by jekyll-scholar.

2026

A Tri-System Risk Model for Human–AI Interaction: Bias Migration and Flash-Risk Dynamics

Seonghyun

2026

Working paper, in preparation

Abs

A simulation model in which interaction itself is the locus of risk: an agent’s effective bias is a stochastic mixture of a human-anchored term and an AI forecaster term, and a loss-aware update governs the tendency to consult the AI. Studies bias migration and flash-risk regimes under varying accuracy, trust, and cognitive load.