安全なGenAIおよびエージェントAIシステムの構築

Meet your instructor, Derek Fisher, and get an overview of the course: the current state of AI adoption, the security frameworks emerging for AI, and why AI security needs a different approach from traditional cybersecurity.

Look at the scale of AI investment and adoption, the main types of GenAI architectures, and why traditional security tools fall short when the attack is semantic rather than syntactic, including new gaps in the AI supply chain.

An introduction to the OWASP GenAI Security Project, what it covers and who contributes to it, followed by a walk through the nine phases of the LLMSecOps framework used throughout the course.

An overview of the NIST AI Risk Management Framework and its risk-based approach, covering its main components: the Core, the Playbook, the Crosswalks and the Roadmap.

Understand the four core functions of the NIST AI RMF (Govern, Map, Measure and Manage) and how they work as a continuous cycle with governance at the center.

Turn the NIST AI RMF into action with the Playbook's guidance for each function, from defining roles and building an AI inventory to measuring risk and safe decommissioning, and review the seven characteristics of trustworthy AI.

Get to know MITRE ATLAS, the ATT&CK-style knowledge base of adversary tactics, techniques, mitigations and case studies for AI and machine learning systems, and how it gives teams a common language for AI threats.

See how the NIST Cyber AI Profile extends the six CSF 2.0 functions with AI-specific considerations across its Secure, Defend and Thwart focus areas, so AI security builds on the program you already run.

A whiteboard walkthrough of how the NIST AI RMF, the OWASP GenAI Security Project and MITRE ATLAS layer together, following one example from vulnerability to threat technique to managed risk.

Survey the global AI regulatory landscape, including the EU AI Act's risk classification tiers, US federal direction and OMB requirements, and the growing patchwork of state AI laws.

20 questions on the AI security landscape, the OWASP GenAI Security Project and LLMSecOps, the NIST AI RMF and Cyber AI Profile, MITRE ATLAS, and EU and US AI regulation. Each answer includes an explanation.