This study investigates the evolving intersection of explainable artificial intelligence (XAI) and the financial sector. It explores how machine learning models’ transparency and interpretability shape decision-making processes in areas such as credit scoring, risk assessment, and market prediction. While conventional AI methods often function as opaque black boxes, XAI offers a solution to promote model accountability, fairness, and trust, which are critical factors in highly regulated and risk-sensitive financial environments. Drawing on 90 peer-reviewed articles retrieved from Scopus and Web of Science, this research applies both co-word analysis and BERTopic modeling to uncover major research themes and semantic structures within the relevant literature. The co-word network reveals distinct thematic clusters related to explainable credit risk models, interpretability in financial forecasting, and the integration of environmental, social, and governance (ESG) factors into AI-driven financial analysis. Meanwhile, topic modeling uncovers additional topics, including SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME)-based explainability techniques, human-AI collaboration in decision-making, and explainable deep learning for asset pricing and sovereign risk analysis. A temporal analysis of publication trends indicates increasing scholarly attention to transparency in AI in the wake of regulatory pressure and ethical considerations in finance. The findings point to a growing emphasis on embedding interpretability into AI models to support fairness, regulatory compliance, and better-informed financial judgments. This study contributes to both academic discourse and practical application by offering a detailed map of current XAI research in finance and identifying key directions for future inquiry and technological development.