Episode 43: Are You Talking to AI? Deepfakes, GANs, and the AI Security Menu

We dive deep into the dark side of Generative AI, exploring how attackers are leveraging Generative Adversarial Networks (GANs) to bypass authentication, and why defining "AI Security" requires much more than just a buzzword.What You’ll Learn:◈ Weaponized AI: How deepfakes, voice cloning, and AI note-takers are disrupting the identity perimeter and creating new insider threats.◈ The User Burden: Why end-users inevitably bear the brunt of corporate security failures and the rising risk of synthetic identities.◈ Zero Trust for AI: Strategies for isolating anomalies, maintaining visibility, and managing the risks of agentic AI.◈ The AI Security Menu: Breaking down the 5 distinct types of AI security—from consuming commercial LLMs to building proprietary models.Key Moments:01:36 ➔ The Deepfake Threat: Why your voice could be used to bypass security protocols.09:59 ➔ Zero Trust vs. AI: The critical importance of visibility, discoverability, and continuous monitoring.12:26 ➔ Third-Party Risk Management for LLMs: Moving beyond basic vendor checks to test for prompt injection and token manipulation.14:48 ➔ The AI Security Menu: How to classify your AI risk footprint.08:04 ➔ THE QUOTE: "Even if you don't think you have an online presence, you do in some way."Subscribe to our LinkedIn to never miss news, updates, and quizzes to earn digital badges.https://ztjourney.comLinkedIn YouTubeDisclaimer: The views expressed are those of the speakers.