Goblins in the Machine: Threat Modeling AI Systems & Their Hidden Attack Vectors
NetSPI Speakers
Date:
April 30, 2026
Location: Toronto, ON, Canada
Resources
As artificial intelligence rapidly transforms the technology landscape, NetSPI Director of Solutions Architecture Mark Davenport demonstrates how established security practices and AI threat modeling can help organizations move quickly without sacrificing security. Drawing from real-world security testing and engineering principles Mark explores:
The Three Pillars of AI Security Testing
1. Model security
2. Application security
3. Holistic pipeline threat modeling
This iCONIC 2026 talk also details how security teams can apply proven practices to modern AI environments by decomposing complex cloud architectures, defining trust zones, and creating comprehensive threat traceability matrices.