The Role of AI-Based Decision Support in Financial Risk Management and Anomaly Detection

Abstract

This study has examined the role of artificial intelligence-based decision support in financial risk management and anomaly detection within selected United States banking, capital markets, and financial technology contexts. The central problem has been that financial institutions face rising volumes, velocity, and complexity of transactions and exposures, in which fraud, market anomalies, credit deterioration, and operational risk events can develop and propagate faster than conventional rule-based monitoring and periodic manual review can detect them, while stand-alone automation is constrained by opacity, alert fatigue from high false-positive rates, regulatory demands for explainability, and the need for accountable human oversight. The purpose of the study has been to evaluate whether model and data quality integration, anomaly detection accuracy and coverage, explainability, interpretability and transparency, human oversight, trust and governance, and real-time monitoring and scalability significantly improve risk management and performance effectiveness when AI-based decision support is deployed as a collaborative capability rather than as an autonomous black box. A quantitative, cross-sectional, case-based research design has been used, and data have been collected from risk managers, quantitative analysts, fraud and financial-crime specialists, model-risk and compliance officers, and financial data scientists engaged in risk monitoring and decision-making. Out of 277 distributed questionnaires, 236 valid responses have been retained, producing an 85.2% valid response rate. The analysis plan has included descriptive statistics, Cronbach’s alpha reliability testing, Pearson correlation analysis, multiple regression modeling, hypothesis testing, a prototype anomaly-detection evaluation, and Python-driven analytical workflow validation. The findings have shown strong agreement across the constructs, with model and data quality integration recording the highest mean score of 4.23, followed by anomaly detection accuracy and coverage at 4.16, explainability, interpretability and transparency at 4.07, risk management and performance effectiveness at 4.02, real-time monitoring and scalability at 3.96, and human oversight, trust and governance at 3.89. Reliability has been confirmed through Cronbach’s alpha values ranging from .82 to .91, with overall reliability of .93. Correlation results have shown significant positive relationships between risk management effectiveness and anomaly detection accuracy (r = .69), model and data quality integration (r = .68), explainability (r = .64), human oversight and governance (r = .60), and real-time monitoring (r = .58), all at p < .001. The regression model has been significant, F(5, 230) = 54.02, p < .001, explaining 54.0% of the variance. The study implies that AI-based decision support can meaningfully strengthen financial risk management and anomaly detection when data quality, detection accuracy, explainability, governed human oversight, and real-time scalability are jointly present.