Hands-On AI System Security: Attacks on ML Models & Cyber Defense

Artificial Intelligence is transforming modern technology, from healthcare and finance to autonomous systems and cybersecurity. However, as AI systems become more powerful and widely deployed, they are also becoming prime targets for sophisticated cyberattacks. Machine learning models can be manipulated, poisoned, stolen, or deceived—creating serious security and privacy risks for organizations worldwide. Hands-On AI System Security: Attacks on ML Models & Cyber Defense provides a practical and research-driven exploration of the rapidly evolving field of AI security. This book guides readers through real-world attacks against machine learning systems, including adversarial attacks, data poisoning, model evasion, prompt injection, model inversion, model extraction, deepfake manipulation, and AI-driven cyber threats. Alongside attack methodologies, it presents effective defense mechanisms, secure AI development practices, threat detection strategies, explainable AI security techniques, and modern cyber defense frameworks. The book offers detailed coverage of secure machine learning pipelines, AI risk assessment, adversarial training, federated learning security, cloud AI protection, and AI governance principles. Readers will explore how attackers exploit vulnerabilities in neural networks, large language models (LLMs), computer vision systems, and intelligent automation platforms, while also learning practical strategies to mitigate these threats using defensive AI techniques and security monitoring frameworks. Key topics covered in this book include: Fundamentals of AI and Machine Learning Security Adversarial Machine Learning Attacks Data Poisoning and Model Manipulation Prompt Injection and LLM Security AI Malware and Automated Cyber Threats Secure AI Model Deployment and Monitoring