Securing AI Agents: Foundations, Frameworks, and Real-World Deployment
暫譯: 保護 AI 代理：基礎、框架與實際部署
Huang, Ken, Hughes, Chris
- 出版商: Springer
- 出版日期: 2025-10-02
- 售價: $3,330
- 貴賓價: 9.5 折 $3,164
- 語言: 英文
- 頁數: 373
- 裝訂: Hardcover - also called cloth, retail trade, or trade
- ISBN: 3032021294
- ISBN-13: 9783032021298
-
相關分類:
Reinforcement
海外代購書籍(需單獨結帳)
商品描述
This book focuses on agentic AI security, providing a comprehensive guide to the theoretical foundations and practical techniques required to secure the increasingly prevalent AI agent systems. It examines the security challenges posed by multi-agent environments and presents real-world examples of open-source frameworks and commercial solutions to mitigate these risks. It answers key questions, including how to conduct threat modeling for agentic AI systems, how to secure communication and identity within multi-agent environments, and how to leverage open-source frameworks and commercial solutions for effective security.
The book features dedicated chapters on agentic AI threat modeling, identity security, communication security in MAS (Multi-Agent Systems), red teaming, AI agents life cycle security, capability and security benchmarking using GAIA and AIR frameworks, Reinforcement Learning (RL) and security, secure agentic AI deployment strategies, innovative open source security frameworks (Cloud Security Alliance and OWASP examples), and case studies of commercial startups addressing agentic AI security challenges. It also explores the unique threat landscape of agentic AI, the challenges of securing communication and identity within multi-agent systems, and the practical application of security benchmarks and open-source frameworks.
As such, the book equips cybersecurity professionals, AI developers, and researchers with the knowledge and tools to mitigate the unique security risks associated with autonomous agents and multi-agent systems.
商品描述(中文翻譯)
本書專注於代理式人工智慧（agentic AI）安全，提供了全面的指南，涵蓋了保護日益普及的人工智慧代理系統所需的理論基礎和實用技術。它探討了多代理環境所帶來的安全挑戰，並提供了開源框架和商業解決方案的實際案例，以減輕這些風險。書中回答了幾個關鍵問題，包括如何為代理式人工智慧系統進行威脅建模、如何在多代理環境中保護通信和身份，以及如何利用開源框架和商業解決方案來實現有效的安全性。
本書包含專門章節，涵蓋代理式人工智慧威脅建模、身份安全、MAS（多代理系統）中的通信安全、紅隊測試、人工智慧代理的生命週期安全、使用GAIA和AIR框架的能力與安全基準、強化學習（Reinforcement Learning, RL）與安全性、安全的代理式人工智慧部署策略、創新的開源安全框架（如雲安全聯盟和OWASP的例子），以及針對代理式人工智慧安全挑戰的商業初創公司的案例研究。它還探討了代理式人工智慧獨特的威脅環境、多代理系統中通信和身份安全的挑戰，以及安全基準和開源框架的實際應用。
因此，本書為網路安全專業人員、人工智慧開發者和研究人員提供了減輕與自主代理和多代理系統相關的獨特安全風險所需的知識和工具。
作者簡介
Ken Huang is a globally recognized expert in AI and Web3 security, a prolific author, and a leading figure in shaping industry standards. He is the CEO and Chief AI Officer (CAIO) of DistributedApps.ai, specializing in generative AI training and consulting. Ken is deeply involved in driving the development of secure AI systems. He serves as a Research Fellow and Co-Chair of the AI Safety Working Groups at the Cloud Security Alliance (CSA), leading agentic AI initiatives. He is also a Co-Chair of the AI STR Working Group at the World Digital Technology Academy under the UN Framework, and a core contributor to the OWASP GenAI project, focusing on agentic AI security. His expertise extends to his contributions to OWASP's Top 10 Risks report for LLM Applications and his active participation in the NIST Generative AI Public Working Group. He is also a member of the Open AI forum. He has authored and edited numerous influential books in this field. His co-authored book, "Blockchain and Web3: Building the Cryptocurrency, Privacy, and Security Foundations of the Metaverse" (Wiley, 2023), was recognized as a must-read by TechTarget. As a sought-after speaker, Ken has presented at prestigious global forums, including Davos WEF, ACM, IEEE, CSA AI Summit, the Depository Trust & Clearing Corporation (DTCC), and World Bank conferences.
Chris Hughes is the Co-founder and CEO of Aquia, a cybersecurity consulting firm dedicated to securing digital transformation initiatives. With nearly two decades of experience in IT and cybersecurity, Chris leads Aquia with a strong commitment to innovation, security, and impact. He previously served as a Cyber Innovation Fellow (CIF) at the Cybersecurity and Infrastructure Security Agency (CISA), where he focused on advancing software supply chain security. Chris also advises several emerging technology startups in areas including Software Composition Analysis (SCA), Kubernetes Security, Non-Human Identities (NHI), and AI Security. In the private sector, Chris has worked as a consultant and currently serves as an adjunct professor for cybersecurity master's programs at the University of Maryland Global Campus. He actively contributes to the cybersecurity community through his involvement in industry groups such as the Cloud Security Alliance's Incident Response and SaaS Security Working Groups, and serves as the Membership Chair for Cloud Security Alliance D.C..
作者簡介(中文翻譯)
Ken Huang 是全球公認的人工智慧 (AI) 和 Web3 安全專家， prolific author，並且在塑造行業標準方面扮演著重要角色。他是 DistributedApps.ai 的執行長及首席 AI 官 (CAIO)，專注於生成式 AI 的訓練和諮詢。Ken 深度參與推動安全 AI 系統的發展。他擔任雲安全聯盟 (CSA) 的 AI 安全工作組的研究員及共同主席，領導代理 AI 的倡議。他也是聯合國框架下的世界數位科技學院 AI STR 工作組的共同主席，並且是 OWASP GenAI 項目的核心貢獻者，專注於代理 AI 的安全性。他的專業知識延伸至他對 OWASP LLM 應用程序十大風險報告的貢獻，以及他積極參與 NIST 生成式 AI 公共工作組。他也是 Open AI 論壇的成員。他在這個領域撰寫和編輯了多本有影響力的書籍。他共同撰寫的書籍《Blockchain and Web3: Building the Cryptocurrency, Privacy, and Security Foundations of the Metaverse》（Wiley, 2023）被 TechTarget 認可為必讀書籍。作為一位受歡迎的演講者，Ken 曾在包括達沃斯世界經濟論壇 (WEF)、ACM、IEEE、CSA AI 峰會、存託信託與清算公司 (DTCC) 和世界銀行會議等多個全球知名論壇上發表演講。
Chris Hughes 是 Aquia 的共同創辦人及執行長，這是一家專注於保障數位轉型計劃的網路安全諮詢公司。擁有近二十年的 IT 和網路安全經驗，Chris 以強烈的創新、安全和影響力的承諾領導 Aquia。他曾擔任網路安全和基礎設施安全局 (CISA) 的網路創新研究員 (CIF)，專注於推進軟體供應鏈安全。Chris 也為幾家新興科技初創公司提供諮詢，涵蓋軟體組成分析 (SCA)、Kubernetes 安全、非人類身份 (NHI) 和 AI 安全等領域。在私營部門，Chris 曾擔任顧問，並目前擔任馬里蘭大學全球校區網路安全碩士課程的兼任教授。他通過參與行業團體，如雲安全聯盟的事件響應和 SaaS 安全工作組，積極為網路安全社群做出貢獻，並擔任雲安全聯盟 D.C. 的會員主席。
