THE CREDIT BLACK BOX: EXPLAINABLE AI AND ALGORITHMIC FAIRNESS IN DIGITAL LENDING – A SAFEGUARD FRAMEWORK FOR NEW-TO-CREDIT BORROWERS IN INDIA
Scholar Journal of Commerce and Management · 2026
DOI: 10.65219/sjcm.20260203017 ↗Kurz gefasst
Fachartikel in der Zeitschrift Scholar Journal of Commerce and Management, 2026. Verfasst von S. Bavatharini und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Digital lending has transformed credit assessment in India by enabling faster, data-driven lending decisions through artificial intelligence (AI) and alternative sources of information. However, when automated credit decisions are produced by complex machine-learning systems, borrowers may receive a decision without understanding why their application was approved, rejected or assigned particular loan terms. This “credit black box” is especially important for new-to-credit (NTC) borrowers, who have limited or no formal credit history and may therefore be assessed using alternative digital signals. This study examines whether explainable artificial intelligence (XAI) can contribute to fairer and more inclusive digital lending for NTC borrowers in India. It adopts a descriptive and analytical research design based on secondary sources, including Reserve Bank of India reports and guidelines, recent credit-market evidence and published research on explainable machine learning, algorithmic fairness and responsible credit assessment. The analysis focuses on four dimensions: the transparency of automated credit decisions, responsible use of borrower data, fairness of algorithmic assessment, and access to meaningful explanation and review. On this basis, the study proposes an Explainability-to-Inclusion Framework that connects decision explanation with data governance, fairness auditing, human review and accessible borrower recourse. The analysis indicates that XAI can improve the visibility and auditability of the factors influencing automated credit decisions, but explainability alone cannot guarantee fairness: if alternative data contain socio-economic or behavioural proxies, a transparent model may still reproduce existing forms of exclusion. Responsible digital lending therefore requires explainability to be supported by consent-based data use, representative data, fairness audits, human oversight and mechanisms for correcting or challenging decisions. The study concludes that XAI should be viewed not merely as a technical feature but as a potential borrower-protection mechanism that strengthens transparency, accountability and responsible financial inclusion.
Zitieren
APA
Bavatharini, S., Ranjani, M. J., Hasna, S., Jasmitha, S., Keerthiga, C., & Kavimalar, B. (2026). THE CREDIT BLACK BOX: EXPLAINABLE AI AND ALGORITHMIC FAIRNESS IN DIGITAL LENDING – A SAFEGUARD FRAMEWORK FOR NEW-TO-CREDIT BORROWERS IN INDIA. Scholar Journal of Commerce and Management. https://doi.org/10.65219/sjcm.20260203017
Vancouver
Bavatharini S, Ranjani MJ, Hasna S, Jasmitha S, Keerthiga C, Kavimalar B. THE CREDIT BLACK BOX: EXPLAINABLE AI AND ALGORITHMIC FAIRNESS IN DIGITAL LENDING – A SAFEGUARD FRAMEWORK FOR NEW-TO-CREDIT BORROWERS IN INDIA. Scholar Journal of Commerce and Management. 2026. doi:10.65219/sjcm.20260203017
BibTeX
@article{Bavatharini2026,
author = {Bavatharini, S. and Ranjani, M. Jeeva and Hasna, S. and Jasmitha, S. and Keerthiga, C. and Kavimalar, B.},
title = {THE CREDIT BLACK BOX: EXPLAINABLE AI AND ALGORITHMIC FAIRNESS IN DIGITAL LENDING – A SAFEGUARD FRAMEWORK FOR NEW-TO-CREDIT BORROWERS IN INDIA},
journal = {Scholar Journal of Commerce and Management},
year = {2026},
doi = {10.65219/sjcm.20260203017}
}
EUVIMED