Regulatory challenges and compliance in federated learning (FL) for financial applications

View at publisher Summarise with VIRAT
Authors: K. Logeswaran, S. Savitha, P. Suresh, K. R. Prasanna Kumar, A. P. Ponselva Kumar, A.S. Jayasurya
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The financial sector is increasingly turning toward artificial intelligence (AI) for applications such as fraud detection, credit scoring, and risk management. But that makes it contrary to the regulatory environment. New data protection laws such as the General Data Protection Regulation (GDPR) in Europe and the Digital Personal Data Protection Act (DPDPA) in India impose stringent conditions on data residency, minimization, and sovereignty. This chapter argues that traditional centralized AI systems which require sensitive data to be collected for processing at one site simply do not sit well with these legal requirements, thereby creating massive compliance risks for financial institutions. By way of an extensive architectural study and practical application, this chapter demonstrates that the very basic functions of a traditional AI system tend to contravene prohibitions on cross-border transfers of data. Instead, we propose Federated Learning (FL) as a “compliance-by-design” solution that solves this sticking point. In other words, by inverting the discredited approach—and bringing the algorithm to the data rather than the other way around—FL ensures that practitioners in different institutions and jurisdictions collaborate on model training without sharing raw data. Only aggregated and anonymized updates on the model are sent—inherently complying with certain data residency and data minimization principles. Besides advocating for FL as a core compliant innovation pathway, this chapter also touches on a number of regulatory uncertainties and other potential issues arising from this technology, such as liability, model security, and a need for industry-wide standards. To this end, the chapter clearly states that the adoption of privacy-preserving technologies such as FL has become integral.

Fields of Study & Subjects

Privacy-Preserving Technologies in Data Digitalization Law and Regulation Ethics and Social Impacts of AI

Publication details

Journal / source
CRC Press eBooks
Pages
1–18
Publication date
25 March 2026
Language
EN
Document type
Book chapter