AI Security has become one of the fastest-moving areas in cybersecurity, and understanding it deeply — not just superficially — is now a core expectation for security engineers, ML practitioners, and IT leaders working with AI systems. This course delivers 600 rigorously researched, scenario-based practice questions across six full-length tests, each modeled on how AI security concepts actually show up in real interviews, red-team engagements, and day-to-day governance decisions. The course is grounded in primary sources: NIST AI 100-2e2025 (the foundational adversarial machine learning taxonomy), MITRE ATLAS (real-world adversary tactics and techniques), the 2025 OWASP Top 10 for LLM Applications, and the OWASP Top 10 for Agentic Applications 2026. It covers the full arc of AI security — evasion and poisoning attacks against predictive AI, privacy attacks like model extraction and membership inference, prompt injection and LLM-specific risks, agentic AI and multi-agent security, and the governance practices that mature AI-security programs actually rely on. Every single question includes a detailed explanation for every answer option — not just the correct one — so you understand exactly why each choice is right or wrong. Questions build progressively within and across tests, using realistic organizational scenarios spanning healthcare, finance, government, critical infrastructure, and more. Sample question from Test 2 (Evasion & Poisoning Attacks): "Why is the Fast Gradient Sign Method considered a foundational but genuinely limited evasion technique compared to an iterative approach like Projected Gradient Descent?" — with four fully explained answer choices. Whether you're preparing for a security interview, building toward an AI security role, or deepening your understanding of this critical field, this course provides thorough, honest, and technically accurate preparation.