Reinforcing Digital Banking Onboarding with Generative AI Fraud Detection

The following paper contains a deep learning-based solution that was implemented in Python and aims to identify document forgery when undergoing digital on boarding. To grant wide applicability to the study, three different datasets are used including passport images, drivers license images, and bank statement images as a part of the overarching applicability to common identity and financial documents utilized in e-KYC processes. Extensive preprocessing, transformation methods and visualisations were used to facilitate sound training and explain ability. The findings indicate high levels of convergence and learning in all the datasets used which validates the reliability of the model in detecting forgery of multi-document. On the whole, it is demonstrated in the paper that Generative AI applied to computer vision can significantly boost e-KYC-related systems through the elimination of fraud and enhancement of digital identity verification in the financial sector and the management of other services.