GENERATIVE AI IN REAL-TIME E-COMMERCE FRAUD DETECTION: A COMPARATIVE AND ETHICAL ANALYSIS

Abstract

The rapid growth of e-commerce platforms has significantly increased the risk of online fraud, creating the need for advanced and intelligent fraud detection systems. This paper explores the application of Generative Artificial Intelligence models, including Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and a hybrid GAN–VAE framework, for real-time fraud detection in e-commerce environments. GANs are highly effective in generating synthetic fraudulent transaction data, which helps improve the training of fraud detection systems. Meanwhile, VAEs are capable of identifying complex and hidden patterns within transaction data, enabling the detection of anomalous and fraudulent activities. The proposed hybrid approach integrates the strengths of both GANs and VAEs to achieve higher detection accuracy, improved adaptability, and enhanced performance in realtime scenarios. In addition to technical evaluation, the study also addresses important ethical considerations, including data privacy, algorithmic bias, and responsible AI usage. The findings highlight the potential of generative AI techniques to significantly enhance fraud detection frameworks while ensuring ethical and secure implementation in modern e-commerce systems.