고령자 보이스피싱 예방을 위한 인공지능 기반 음성합성 훈련 플랫폼 개발 = Development of an AI-Based Voice Synthesis Training Platform for Voice Phishing Prevention on Older Adults

다국어 초록 (Multilingual Abstract)

Voice phishing has caused serious economic and psychological damage, particularly targeting older adults. With recent advances in AI-based voice synthesis technologies, voice phishing attacks that imitate real individuals have become increasingly sophisticated, revealing the limitations of conventional prevention education based primarily on information delivery. This study designs and implements an integrated platform for voice phishing prevention training for older adults using an AI-based voice synthesis learning model and a voice data collection system. The proposed system enables older adults to collect voice data directly through a mobile application and generates personalized voice-based training content using a voice synthesis model that operates effectively in low-resource data environments. In addition, the training platform is designed based on a three-stage learning flow of “listening–judgment–feedback,” allowing older adults to improve their awareness of voice phishing scenarios through repeated experiential training. System implementation and functional verification confirm that the entire pipeline—from voice data collection and model training to voice synthesis and training content delivery—operates stably as intended. This study verifies the technical feasibility of a voice synthesis–based voice phishing prevention training system using a limited amount of older adult voice data and is expected to serve as foundational work for developing experiential prevention education systems for older adults in the future.