Integrating AI security into cybersecurity programs can be challenging due to limited budgets, IT restrictions, and unpredictable API costs. This interactive session provides instructors with a robust, repeatable framework for teaching AI hardening. In this webinar, the presenter will demonstrate how to use Docker Desktop to run local, open-source Large Language Models (LLMs) on student machines, establishing a safe, isolated sandbox. We will also show how to route and secure this local traffic through a security gateway proxy to inspect, log, and filter malicious payloads. Drawing from open-source GitHub materials and proven classroom setups, attendees will receive ready-to-use Docker configurations, step-by-step lab scenarios for vulnerabilities like prompt injection, and effective pedagogical strategies for teaching AI defense.
Joshua Wortz combines leadership experience with advanced cybersecurity expertise to deliver engaging, real-world instruction. He earned a Bachelor of Arts in Pastoral Ministry from the University of Valley Forge and is completing a Master of Science in Cybersecurity Operations and Leadership at the University of San Diego (expected 2026). Joshua specializes in the NIST Cybersecurity Framework (CSF), red/blue team operations, identity-based threats, cloud security, and open-source intelligence (OSINT) and holds the following professional certifications:
The National Cybersecurity Training & Education (NCyTE) Center is an advanced technological education National Science Foundation grant award #2054724.
Office: 360.383.3175Email: ncyte@whatcom.edu
Address:237 W Kellogg RoadBellingham, WA 98226
