AI-Powered Voice and Video Spoofing Now Within Reach of Midtier Threat Actors

Voice phishing and deepfake video attacks have crossed a capability threshold. Real-time video generation and low-latency voice cloning APIs are now accessible to moderately resourced threat actors—a shift from techniques once requiring significant compute and production time. Spear phishing augmented by LLMs compounds the risk. This shift in the cost and skill floor for social engineering attacks is driving a new category of enterprise security training focused explicitly on these AI-native threat vectors.