Approaches to the Detection of Deepfake in the Financial Organization’s Activities

Ch. 1. P. 3–7.

Belov A. V., Fedotov G.

In recent years, significant progress has been observed as content generated using AI technologies. In addition, tools regularly appear with which scammers can create a realistic fake content. Deepfake detection methods are currently actively used in the activities of financial organizations. With their help, a departments within financial organizations responsible for IT Security identify cases of fraud, protect customers and ensure the safety of digital transactions. To generate deepfake, multimodal models are used to form fake dynamic video images with sound. The work investigates generative models in the Face Synthesis task, as well as methods for detecting deepfakes created using models of this class. The analysis of the datasets used to detect deepfakes is given. Recent studies have demonstrated the effectiveness of these approaches in controlled settings

Keywords: cybersecurity, Artificial Intelligence, deepfakes detection, deepfakes attacks, fintech security, deep learning

In book

Proceedings of the 2025 INTERNATIONAL CONFERENCE "QUALITY MANAGEMENT, DIGITAL SECURITY, INFORMATION TECHNOLOGIES" (2025 QM&DS&IT)

IEEE, 2025.