Security and Privacy in AI

Overview

Security and Privacy in AI equips participants with essential knowledge and skills to navigate the complex landscape of AI security and privacy. The course runs over two comprehensive days. It addresses the growing need for professionals who can understand and mitigate AI risks while maximizing its potential benefits. This is a gap most organizations feel acutely as they move AI pilots into production without a matching security or privacy program.

Day one opens with an AI security and privacy overview that weighs the value AI brings to an organization against its inherent risks. Participants then study the NIST Artificial Intelligence Risk Management Framework, which provides a structured approach to identifying and managing AI-related risks. The course then turns to privacy considerations specific to AI systems, from training data handling to model output exposure. The day closes with the ISO/IEC CD 27090 guidance, which catalogs security threats and potential failures across AI implementations.

Day two turns to practical defense, starting with LLM vulnerabilities and the mitigations that blunt them. Attendees explore observability techniques for tracking AI system behavior in production and apply secure coding practices tailored to AI development. They also configure access control mechanisms that govern who can query, retrain, or modify AI systems. Lecture and demos are paired with hands-on labs throughout, so participants practice each mitigation rather than just reading about it. By the end, attendees have a comprehensive understanding of AI security and privacy issues. This lets them reason about risks and implement effective mitigation strategies in their organizations.

Duration

2 days