Real-time voice modulation detection: Protecting against AI-enabled ransomware call scams

Sharma, Manas (2025) Real-time voice modulation detection: Protecting against AI-enabled ransomware call scams. World Journal of Advanced Engineering Technology and Sciences, 15 (2). pp. 1005-1015. ISSN 2582-8266

This article addresses the growing threat of AI-enabled voice modulation scams by developing a comprehensive framework for real-time detection on mobile devices. The article examines current detection methodologies including machine learning classification, statistical anomaly detection, watermarking, model fingerprinting, and adversarial frameworks. Technical challenges are analyzed across acoustic feature extraction, temporal inconsistency identification, prosodic pattern recognition, real-time processing constraints, and differentiation between legitimate and fraudulent voice alterations. The article presents a client-side implementation architecture optimized for resource constraints, privacy preservation, telecommunications infrastructure integration, and user experience considerations. Experimental evaluation demonstrates significant performance advantages over existing systems, with the proposed approach achieving high accuracy while maintaining computational efficiency and resilience against adversarial attacks. This article concludes by identifying current limitations and outlining promising future research directions to enhance detection capabilities while preserving trust in voice communication.