Uczenie maszynowe w procesie walidacji modeli IRB. Regulacyjne aspekty wykorzystania z perspektywy stanowisk EBA, ECB i KNF

Krzysztof Iwanicki

doi:10.26354/bb.4.2.103.2026

Published: 2026-10-30

Vol. 103 No. 2 (2026)

Machine learning in the IRB model validation process. Regulatory aspects of its use from the perspective of the EBA, the ECB and the KNF

Krzysztof Iwanicki

Section: Problems and Opinions

https://doi.org/10.26354/bb.4.2.103.2026

Abstract

Credit risk is one of the fundamental risk categories inherent to banking activities. Identifying and measuring it is important for the safety of individual credit institutions and the stability of the entire financial sector. In this regard, IRB models play a significant role; they are a key element of the credit risk management system that allows banks to use internal estimates of risk parameters. This article analyzes the theoretical and regulatory basis for using artificial intelligence systems in the IRB model validation process. The article concludes that artificial intelligence systems can be used in the IRB model validation process, but their role should remain auxiliary.

Keywords:

IRB models , IRB model validation , AI systems , artificial intelligence , credit risk

JEL Codes

G21, G28, G32, G52, K23

Download files

pdf (Język Polski) ENG_pdf

Citation rules

Iwanicki, K. (2026). Machine learning in the IRB model validation process. Regulatory aspects of its use from the perspective of the EBA, the ECB and the KNF. Safe Bank, 103(2), 76–86. https://doi.org/10.26354/bb.4.2.103.2026

Cited by / Share