In modern banking, ensuring that new payment systems operate accurately and securely requires extensive testing. However, testing with real-world data introduces privacy risks, and synthetic data offers a promising alternative. This paper explores the potential of Generative AI for producing realistic, privacy‑compliant synthetic transaction data. The proposed approach addresses challenges such as data privacy, diverse dataset creation, and the ability to simulate rare or edge-case scenarios—thus enhancing the robustness of payment systems.
Keywords: Privacy, Synthetic data, generative AI, machine learning, payment systems
