This book is an essential guide for anyone looking to build secure and resilient generative AI applications. It provides actionable insights into identifying vulnerabilities and preventing attacks.
Authored by security expert Paul Zenker, the book explores the critical aspects of securing GenAI systems across their lifecycle - design, development, and operation. Through numerous practical examples, detailed illustrations of threat scenarios, and application architectures, readers will gain a comprehensive understanding of potential risks. Step-by-step instructions cover techniques such as prompt injections, jailbreaks, and other attack vectors, equipping readers with the skills to anticipate and mitigate threats effectively.
This resource is ideal for developers, security professionals, AI researchers, and organizations working with generative AI technologies. This book offers the tools and knowledge needed to create systems that can withstand sophisticated attacks in today’s high-stakes digital landscape.
1 Introduction
2 GenAI Basics
3 Why GenAI security is so special
4 Anatomy and attack surface of GenAI applications
5 Security should be planned
6 Data is the gold of the 21st century – and of the attacker
7. Prompt Injections
8. OWASP LLM Top Ten
9. GenAI security testing
10. GenAI and Cybersecurity
11. Don't forget the old lessons
Biography
Paul Zenker is a security consultant and hacker by heart. He started his career at SAP in Pentesting and Red Teaming. With the recent advancements he became very interested in AI Security from an attacker’s perspective and gave many talks on that topic at international conferences. He now works for KPMG consulting clients in the financial sector about cybersecurity and AI security and safety.
