Why Traditional SAST Fails on AI-Generated Code

AI didn't just speed up software development, it changed what application security programs must defend. As AI coding assistants generate code at machine speed and developers integrate AI agents, models, and RAG pipelines into production, traditional scanners generate endless backlogs of unprioritized alerts.

In this webcast, Ben Rieger from Mend.io and Scott Schober from Cybercrime Magazine unpack how to build a modern application security strategy that scales with AI-powered development. Learn how Mend AI, Mend SAST, and Mend SCA help engineering teams eliminate false positives, detect AI agent configuration risks, and bridge the gap between shift-left scanning and active runtime protection.

Watch the complete webcast on demand on BrightTALK: https://www.brighttalk.com/webcast/15811/673651

⏱️ Timestamps:

00:00 - Introduction: The Shift to AI-Powered Development

02:30 - Why Traditional Security Scanners Fail in the AI Era

07:15 - Identifying the Shadow AI Attack Surface: Agents, Prompts, and Models

12:45 - Agentic SAST Triage: Cutting False Positives and Alert Noise

18:20 - AI Agent Configuration Scanning and Risk Reduction

24:50 - Real-Time Runtime Guardrails against Prompt Injection

31:10 - Live Demonstration: Protecting AI Applications from Dev to Runtime

40:00 - Q&A: Practical AppSec Strategies for Security Leaders

#AppSec #AISecurity #DevSecOps #MendAI #SoftwareSecurity #Cybersecurity #SoftwareEngineering