AI-Enhanced Detection and Mitigation of Cybersecurity Threats in Digital Banking

Abstract

The rapid evolution of digital technologies has revolutionized the banking sector, making financial services more accessible and convenient through digital channels. However, this digital transformation has also introduced significant cybersecurity challenges, exposing financial institutions to a range of threats including fraud, data breaches, and malicious attacks. In response to these challenges, this research proposes an advanced AI-enhanced framework designed specifically for detecting and mitigating cybersecurity threats within the realm of digital banking. This research introduces an AI-enhanced framework for detecting and mitigating these threats in digital banking. Our solution includes a web application utilizing machine learning models to predict loan acceptance and detect fraudulent credit card transactions. Employing a Random Forest algorithm for loan prediction and a Support Vector Machine (SVM) for fraud detection, our models achieve precision rates of 92% and 90%, respectively. The system preprocesses datasets, splits them for training and validation, and generates pickle files for real-time predictions via the web application. An adaptive Class Incremental Learning Framework supports continuous improvement in threat detection. This framework enhances digital banking security by enabling real-time monitoring and proactive threat mitigation, thereby safeguarding sensitive financial information and preserving customer trust.