Paper Details

Name: Namrata Kishnani & Alpa Ghosh
Country: India

Year: 2025
Volume: Volume No: 12, Issue No: 2, Year: 2025, (Specials issue-1)
Page Number: 143-151
DOI: https://doi.org/10.5281/zenodo.17839532

ABSTRACT:
As digital financial services rapidly proliferate across India, cyber threats and financial
fraudpose growing challenges to secure financial infrastructure. This research examines how
artificialintelligence (AI) can be effectively deployed by financial institutions and regulators
in India toprevent and mitigate cyber-based financial crimes.This study explores the
implementation ofAI-based tools—such as ensemble learning, neural networks, transformer
models, andexplainable AI (XAI) frameworks—in Indian banks and fintech institutions to
address thegrowing problem of financial fraud.
Using a mixed-methods research approach, this study integrates primary data collected
fromsurveys of customers and banking professionals with transactional datasets from
selectedIndian banks. The study evaluates the performance of various AI models across
metrics likeaccuracy, false positive rate, latency, and privacy compliance. It also examines
stakeholderperceptions regarding trust, explainability, and usability of AI tools in fraud
prevention.Theresults reveal that transformer-based AI models significantly outperform
traditional rule-basedsystems in terms of fraud detection accuracy and reduction in false
positives. Federatedlearning approaches are shown to preserve privacy while integrating
explainable AI toolsimproves analyst trust and speeds up investigation processes. The study
concludes by offeringstrategic recommendations for regulatory adoption and ethical
deployment of AI in India’sfinancial sector.
Keywords: Artificial Intelligence, Financial Cybersecurity, Fraud Detection, Explainable AI,
Indian Banking