International Journal of Scientific Research and Engineering Development

With the recent developments in Artificial Intelligence (AI), realistic synthetic speech can now be produced through voice cloning and neural network-based text-to-speech systems. However, the security threats associated with such technologies are critical in the banking and financial sector. Deepfake voice-based scam calls are a recent form of sophisticated fraud in which attackers pose as bank officials and extract critical information from the victim. The current paper presents a real-time deepfake voice detection system for the banking call scenario. This system can stream audio signals in real-time and use a hybrid Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) model to classify the audio signals as real or fake. This architecture can also generate alerts to prevent financial fraud. Experimental results on benchmark datasets have proven the effectiveness of the proposed system. The proposed system can greatly enhance communication security and can be used as a robust defense mechanism against AI-based voice cloning technologies scams.

Dr.K.Karuppasamy, R.Thenmalar, S.Jayaraman, P.Gokul, R.Janarthan, M.Ajay Makeshwaran,"Deepfake Voice Based Scam Detection in Banking Calls Using AI and ML" International Journal of Scientific Research and Engineering Development, V9(2): Page(2292-2298) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.