Rethinking AppSec for the AI Era

AI-driven systems are reshaping the application security landscape faster than most organizations can adapt. In this episode, Bar-El Tayouri, Head of AI Security at Mend.io, breaks down how multi-agent architectures, fuzzy interfaces, and emerging model components are expanding the attack surface—and what modern AppSec teams must do to keep up. From advanced threat simulation to rethinking identity and governance, Bar-El outlines a practical path for securing AI at enterprise scale.

This episode covers:

• How modern AppSec must evolve to secure AI agents, MCP servers, system prompts, and model artifacts
• Why AI-driven architectures expand the attack surface with multi-agent systems and fuzzy interfaces
• How Mend discovers hidden and shadow AI usage across an organization
• Dynamic threat simulation techniques, including prompt injection via common file types
• Developer-focused, in-code remediation that reduces AI risk without slowing innovation
• Why identity, access control, separation of duties, and system modeling must be reimagined for AI systems
• Why CISOs need to integrate AI security directly into development workflows instead of relying on blunt-force blocking

Speakers:

• Bar-El Tayouri, Head of Mend AI – Mend.io
• Daniel Miessler, Cybersecurity / AI engineer and founder