Marc Faudi: How Voice Cloning Bypasses Voice Authentication Systems
A DEF CON 32 talk where researcher Marc Faudi demonstrates how neural network–based voice cloning can be used to bypass voice verification systems.
Using a cloned version of his own voice, he simulates attacks that successfully fool authentication systems of several major organizations.
The talk explores tools for creating voice clones, the associated security risks, and possible defenses against such attacks.
Using a cloned version of his own voice, he simulates attacks that successfully fool authentication systems of several major organizations.
The talk explores tools for creating voice clones, the associated security risks, and possible defenses against such attacks.