A Regulatory Governance Framework for AI-Driven Financial Fraud Detection in U.S. Banking: Integrating OCC, SR 11-7, CFPB, and FinCEN Compliance Requirements for Model Development, Validation, and Monitoring Lifecycles

U.S. banks deploying AI for fraud detection face a fragmented compliance landscape across four regulatory frameworks: OCC Bulletin 2011-12, SR 11-7, CFPB AI circular, and FinCEN BSA/SAR requirements. In a new paper submitted to Cogent Business & Management (Taylor & Francis), researcher Mohammad Nasir Uddin presents the Regulatory Governance Framework for AI-Driven Financial Fraud Detection (RGF-AFFD). This three-tier architecture is empirically anchored in a multi-study program using two large transaction datasets: the IEEE-CIS dataset (590,540 transactions) and the ULB benchmark (284,807 transactions). The framework benchmarks six architectures, including an LSTM+XGBoost ensemble, and conducts ablation, temporal drift, SHAP interpretability, and BISG fairness analyses. The Regulatory Digital Twin meta-model translates metrics into four regulator-specific health scores and a composite Regulatory Fitness Index for continuous compliance monitoring.

Key technical results show the LSTM+XGBoost ensemble achieves a ROC-AUC of 0.9289 with an F1 score of 0.6360 and a benefit-cost ratio of 6:1. Notably, XGBoost demonstrates the strongest temporal stability (delta-AUC of -0.0017 compared to -0.0626 for LSTM), making it more reliable for ongoing fraud detection. The RGF-AFFD is the first integrated deployment blueprint that simultaneously satisfies OCC, SR 11-7, CFPB, and FinCEN requirements, supported by a community bank implementation vignette and four evidence-based policy recommendations.