Explainable Artificial Intelligence in the Banking Sector: A Systematic Literature Review
Ntombikayise Betreace Tuge, Nkosikhona Theoren Msweli · Applied Artificial Intelligence · 2026
This systematic literature review investigates the critical implementation gap between the technical promise of Explainable Artificial Intelligence (XAI) and its operational deployment within the banking sector. While the existing literature extensively documents XAI’s potential to enhance transparency, regulatory compliance, and trust, significant barriers prevent its effective integration into banking workflows. Adhering to the PRISMA 2020 methodology, this review synthesizes 78 peer-reviewed studies and industry reports (2021–2025) to analyze why XAI adoption remains challenging despite recognized necessity. Our findings reveal that technical challenges, particularly the accuracy-interpretability trade-off and the “explaining the explanation” problem, are compounded by more fundamental organizational and cultural barriers. These include resistance to change, misalignment between technical teams and business units, and a lack of standardized evaluation frameworks. Crucially, we identify that successful XAI implementation requires addressing socio-technical tensions through interdisciplinary collaboration, context-aware explanation design, and human-centric governance frameworks. This review offers a critical synthesis that moves beyond cataloging XAI techniques to diagnose implementation failures and propose a pathway to bridge the gap between XAI research and banking practice. We conclude that XAI must evolve from a technical add-on to an embedded organizational capability, necessitating structural changes in how banks approach AI governance, stakeholder communication, and ethical oversight.