Security Television Network
Automated AI Security Testing: Find and Fix Security Weaknesses Before Deployment
Last week your AI agent refused a jailbreak. This week you updated the model, changed the prompt, and connected another tool. Is it still secure? The only way to know is to test it again. This demo shows how CrowdStrike is expanding the Falcon platform with Dreadnode, bringing Automated AI Security
About this video
Last week your AI agent refused a jailbreak. This week you updated the model, changed the prompt, and connected another tool. Is it still secure? The only way to know is to test it again.
This demo shows how CrowdStrike is expanding the Falcon platform with Dreadnode, bringing Automated AI Security Testing (AAIST) to the agentic enterprise. Follow Priya as she prepares to launch an AI assistant that handles employee pay and benefits questions, launch 50+ automated attack techniques against it, uncover a critical jailbreak that slipped past its defenses, and validate the fix before it ever reaches production.
You'll see how to: π Describe what to test in plain language, no code required π Run 50+ attack techniques automatically, from jailbreaks to multi-turn manipulation π Replay an entire attack, prompt by prompt, to see exactly where defenses failed π Reclassify findings and watch the risk dashboard update in real time π Map findings to OWASP, MITRE ATLAS, and NIST frameworks for go/no-go decisions
π‘οΈ Learn more about CrowdStrike Automated AI Security Testing: https://cs.link/uuwHf
π£ Connect With Us:
βΊ X: https://twitter.com/CrowdStrike βΊ Instagram: https://www.instagram.com/crowdstrike βΊ LinkedIn: https://www.linkedin.com/company/crowdstrike
π Subscribe to stay updated!
#CrowdStrike #AutomatedAISecurityTesting #Dreadnode
Attribution: CrowdStrike
The interactive STN application appears here when JavaScript is available. Published facts and source links above remain readable without JavaScript.