Explaining the Unexplainable: A Systematic Review of Explainable AI in Finance
Published 11 Nov 2025 · arXiv · Md Talha Mohsin
Overview
The paper provides a comprehensive review of Explainable AI (XAI) applications in finance, emphasizing the balance between accuracy and transparency. It highlights the prevalent use of post-hoc interpretability techniques and the need for integrating financial knowledge with XAI.
Key Insights
- Post-hoc Interpretability Techniques: The study finds a significant reliance on techniques like SHAP and feature importance analysis in financial AI applications.
- Multidisciplinary Approaches: There is a call for combining financial expertise with advanced XAI methods to address current shortcomings.
- Trend Mapping: The review maps trends in XAI research, identifying key topic clusters and methodologies.
BFSI Relevance
- Why Relevant: Transparency in AI models is crucial for regulatory compliance and trust in financial services.
- Primary Sector: Financial Services
- Subsectors: Asset Management, Risk Management
- Actionable Implications: BFSI professionals should advocate for and implement XAI techniques to enhance model transparency and compliance.
researcher peer-reviewed-paper financial-services technology-and-data global