Design of an XAI-Enhanced Machine Learning Audit Model for Ethical, Technical, and Regulatory Compliance in the European Financial Sector

Mazat Ixcol, Jenyffer Noelia (2025) Design of an XAI-Enhanced Machine Learning Audit Model for Ethical, Technical, and Regulatory Compliance in the European Financial Sector. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (4MB) | Preview
Preview	PDF (Configuration Manual) Download (671kB) | Preview

Abstract

The European financial sector is increasingly relying on Artificial Intelligence decision-making, which requires rigorous governance systems and transparent fair practices. The EU Artificial Intelligence Act, GDPR, DORA, EBA Guidelines and CBI expectations have established strict requirements for high-risk AI systems but provide insufficient operational direction resulting in a gap between regulatory expectations and institutional implementation.

This research develops and validates an AI Audit framework that upholds ethical principles, explainability methods (SHAP, LIME, EBM-native explainability) and EU–Irish regulatory requirements, by developing a twelve-pillar system with Disparate Impact, Statistical Parity Difference, Equal Opportunity Difference, fairness assessments, technical validation and automated compliance mapping into a unified auditable workflow.

The Lending Club data is used in a credit-scoring simulation, to generate auditable evidence which follows regulatory requirements for supervisory assessment, results show high predictive accuracy (ROC-AUC > 0.99), complete fairness threshold compliance (DI ≥ 0.80) and the ability to detect operational gaps, especially in Ethical Alignment, Explainability and Legal Basis for Processing from self-reported to evidence-based compliance. The framework enables financial institutions, consultancy firms and regulators to create standardized audit artefacts through a repeatable process which strengthens trust in high-risk AI systems while closing the regulatory gap.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Del Rosal, Victor UNSPECIFIED
Subjects:	H Social Sciences > HG Finance Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence H Social Sciences > HG Finance > Fintech T Technology > T Technology (General) > Information Technology > Fintech Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in FinTech
Depositing User:	Ciara O'Brien
Date Deposited:	20 Aug 2026 11:24
Last Modified:	20 Aug 2026 11:24
URI:	https://norma.ncirl.ie/id/eprint/9576

Actions (login required)

View Item