NIST catalogues four classes of adversarial machine learning attack and OWASP lists ten more for large language models. Most enterprises have inventoried none of them. AI Cybersecurity All-in-One is one reference for the entire AI attack surface, and the defence programme that covers it.
The book maps every part of the surface: training data and poisoning, model theft and extraction, membership inference, prompt injection, insecure plugins, agent misuse and the model supply chain. Each attack is paired with the detection, control and assurance evidence that closes it out. It is built for security leaders, architects, engineers and assurance teams who need a single, coherent defence programme rather than a shelf of point solutions.
MAP. A complete, inventoried view of every model, dataset, endpoint and agent in the estate, with owners and exposure ratings.
HARDEN. Concrete controls against poisoning, extraction, evasion and injection, mapped to ATLAS techniques and CSF 2.0 outcomes.
ASSURE. Testing, monitoring and reporting that convince an auditor, an insurer and a regulator that the programme is real.
Anchored on NIST AI RMF and AI 600-1, the adversarial ML taxonomy AI 100-2, MITRE ATLAS, the OWASP LLM Top 10, ISO/IEC 42001 and the EU AI Act. Inside this volume: 38 chapters in 11 parts, 6 appendices, 66 figures, 81 tables, 522 primary sources and a 311-term index.
Key topics include: generative AI security, LLM security, agentic AI security, adversarial machine learning, data poisoning, model extraction, prompt injection, AI red teaming, AI supply chain security, zero trust for AI, SOC automation, AI assurance and AI regulation.
Professor Kai London is a cybersecurity executive, security architect, author and professional researcher with extensive experience across cybersecurity, critical infrastructure, resilience, governance, risk and technology. His work focuses on complex operational, security and institutional challenges, including cyber resilience, infrastructure protection, evidence, governance, recovery and reconstruction. He writes practitioner-focused professional reference works designed for executives, specialists, researchers and institutional readers.