Episode
AI Vulnerability Management: Why You Can't Patch a Neural Network
- Podcast
- Cloud Security Podcast
- Published
- Jan 13, 2026
- Duration seconds
- 2480
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/cloud-security-podcast-612345/episodes/ai-vulnerability-management-why-you-can-t-patch-a-neural-network/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/cloud-security-podcast-612345/ai-vulnerability-management-why-you-can-t-patch-a-neural-network.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Traditional vulnerability management is simple: find the flaw, patch it, and verify the fix. But what happens when the "asset" is a neural network that has learned something ethically wrong? In this episode, Sapna Paul (Senior Manager at Dayforce) explains why there are no "Patch Tuesdays" for AI models . Sapna breaks down the three critical layers of AI vulnerability management: protecting production models, securing the data layer against poisoning, and monitoring model behavior for technically correct but ethically flawed outcomes . We discuss how to update your risk register to speak the language of business and the essential skills security professionals need to survive in an AI-first world . The conversation also covers practical ways to use AI within your security team to combat alert fatigue , the importance of explainability tools like SHAP and LIME , and how to align with frameworks like the NIST AI RMF and the EU AI Act . Guest Socials - Sapna's Linkedin Podcast Twitter - @CloudSecPod If you want to watch videos of this LIVE STREAMED episode and past episodes - Check out our other Cloud Security Social Channels: - Cloud Security Podcast- Youtube - …