TRACE: A Governance Framework for Measuring Explainability Debt in Production AI Systems
Read the original on arXiv AI →The paper introduces TRACE, a seven-instrument governance framework designed to measure and manage Explainability Debt in production AI systems. Its core metric, the Explainability Debt Score (EDS), tracks the proportion of decisions falling below a defined explainability confidence threshold, while complementary tools such as DART, SHIV, FDE, HVE, AIDE, and ZERO support forecasting, daily validation, drift evaluation, human oversight, audit decisions, and remediation. A twelve‑month case study on a fraud‑detection system demonstrated that TRACE could predict EDS levels months in advance and reveal risk asymmetries invisible to traditional performance metrics.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.