Live at Black Hat: What’s AI Really Capable Of?

This year at Black Hat, the topic of AI was everywhere — from hallway chats to the expo floor. Adam and Cristian took a break from the action for a rare in-person conversation about how adversaries are weaponizing AI, how defenders are using agentic AI, and what we should all be thinking about as AI evolves as an offensive and defensive tool. The AI threat is real, and advanced adversaries in particular are using it to their advantage. They’re improving the wording in social engineering attacks, creating deepfakes in fraudulent job interviews, and targeting victims on a more personal level. FAMOUS CHOLLIMA is an example of one adversary “using it for everything,” the hosts say. SCATTERED SPIDER is another adversary to watch. On the other side, defenders are adopting agentic AI to expedite their response. Adam and Cristian explore the importance of protecting AI workloads, the potential for insider threats with AI models, and the growing need for AI governance and security guardrails. If AI is monitoring security services, they ask, who guards the guardian? Tune in for an in-depth conversation on what AI is really capable of — and stick around for a sneak peek of an upcoming guest episode, where a guest joins to discuss young adversaries moving from online gaming to organized cybercrime.