Building Security into AI - Tutorial

Learn to design AI systems with security as a foundational principle through this comprehensive tutorial that explores the unique security challenges facing artificial intelligence applications. Discover how AI security risks fundamentally differ from traditional software vulnerabilities and master the creation of effective threat models specifically tailored for AI systems. Analyze real-world AI security breaches to understand common attack vectors and failure points. Explore input manipulation techniques including prompt injection, adversarial inputs, and data poisoning attacks that can compromise AI system integrity. Examine data output concerns such as information leakage, model inversion attacks, and unintended data exposure through AI responses. Develop practical defense strategies against emerging AI threats including input validation, output filtering, and secure model deployment practices. Created by Robert Herbig from APIsec University, this tutorial provides hands-on insights into building robust security frameworks for AI applications from the ground up.