Artificial Intelligence in Fraud Detection: Strengthening Identity Data Security in Open Banking

Open banking allows banks to share information with third-party providers in an easy manner through secure channels that result in enhanced usability of financial services as well as innovations. Increased connectivity of the system creates more susceptibility to fraud activities and identity theft. The paper proposes an Artificial Intelligence (AI) detection system for protecting Open Banking transactions and protecting user identity information. Machine learning algorithms such as XGBoost and Random Forest empower the system to identify fraud accurately in real-time with high precision and recall scores. Data imbalance issue needs preprocessing via Synthetic Minority Over-sampling Technique (SMOTE) that improves model sensitivity. The system adopts Representational State Transfer (REST) Application Programming Interface via OAuth2 and Open Identity (OpenID) Connect for secure deployment and employs AES (Advanced Encryption Standard) and RSA (Rivest–Shamir–Adleman) encryption to secure sensitive user information. Multi-factor authentication and tokenization provide additional identity protection. Machine Learning Operations (MLOps) allows the system to undertake continuous learning and model retraining and adaptation for identifying new fraud patterns. Experimental results highlight the superior detection accuracy and AUC (Area under the Curve) of Extreme Gradient Boosting (XGBoost), which determines the effectiveness of the system in real-world financial settings.