As the economy becomes more unstable, there is a growing need for accurate and flexible ways to assess financial risks. Artificial intelligence (AI), with its capacity to simulate uncertainty and optimize decision-making, offers significant opportunities to enhance traditional financial risk models. This study proposes a hybrid predictive modeling framework that integrates AI-driven simulation techniques such as Monte Carlo methods with optimization algorithms including genetic algorithms and reinforcement learning to provide robust, transparent, and adaptive risk assessment. Using a synthetic dataset simulating credit, market, and operational risk scenarios across diverse financial portfolios, the framework was evaluated for accuracy, efficiency, and interpretability. The hybrid model outperformed conventional models in forecasting risk exposure and loss probabilities under both normal and stress conditions. Moreover, incorporating explainable AI components, including SHAP (SHapley Additive exPlanations) values and feature attribution maps, enhanced transparency and stakeholder trust in model outputs. Findings demonstrate that this integrated approach provides significant improvements in predictive performance, model adaptability, and interpretability compared to standalone methods. The study also examines the implications of AI-generated financial forecasts on regulatory compliance, organizational resilience, and ethical risk governance. It concludes with recommendations for the scalable adoption of hybrid AI frameworks in institutional risk environments, emphasizing the importance of transparent modeling and inclusive algorithmic oversight.
Oluwaseun Lamina (Mon,) studied this question.