Bridging the Interpretability Gap: A Comparative Review of Regulatory Trends and Technical Implementations of Explainable AI (XAI) in the Financial Systems of Emerging Economies

As financial institutions transition from traditional statistical models to complex "black-box" AI systems, the demand for transparency has created a critical interpretability gap. This paper provides a comparative review of the regulatory trends and technical implementations of Explainable AI (XAI) in developed (EU/US) versus emerging (India/Vietnam) economies. While developed markets rely on prescriptive mandates like the GDPR and EU AI Act, emerging economies are navigating a spectrum of governance models. Specifically, India maintains an adaptive, principles-based approach through frameworks like the NITI Aayog’s Responsible AI guidelines and RBI’s Digital Lending Guidelines, which prioritise flexibility and sectoral oversight rather than a standalone AI law. In contrast, Vietnam is transitioning toward strict legal boundaries with its Law on Artificial Intelligence (2026), which classifies banking as high-risk. Through a qualitative thematic synthesis, the study identifies that emerging markets often face unique technical and resource hurdles, leading to the adoption of hybrid and heuristic XAI models. The results suggest that XAI in finance is both a technical and a governance challenge, requiring a balance between model performance and the need for trustworthy AI governance. The paper concludes that bridging the digital divide in XAI tools is essential for ensuring financial inclusion and long-term stability in global financial systems.