Real-Time Video and Audio Deep Fake Impersonation - A Security Demonstration
Learn about the concerning rise of real-time deep fake technology through a DEF CON 31 security conference presentation that demonstrates how accessible CEO impersonation has become using video and audio manipulation. Explore the technical components needed for creating convincing deep fakes, including hardware requirements, Deep Face Lab usage, offline and real-time deep fake generation, speech-to-text conversion, and voice manipulation. Follow along with a live demonstration showing how open-source tools and standard GPU hardware can be used to impersonate company executives through video calls. Gain critical insights into this emerging security threat, including social engineering implications and protective countermeasures organizations should consider implementing. The presentation covers both technical implementation details and broader security implications while highlighting how deep fake technology has evolved to enable real-time impersonation attacks.