خرید و دانلود نسخه کامل کتاب زبان اصلی Securing the Thinking Machine: A Practitioner’s Guide to AI Security PDF | دانلود کتاب خارجی

آنتونی رابینز میگه : من در 40 سالگی به جایی رسیدم که برای رسیدن بهش 82 سال زمان لازمه و این رو مدیون کتاب خواندن زیاد هستم.

 
AI systems are no longer tools. They are actors. They make decisions, take actions, and pursue goals on behalf of the organisations that deploy them. And most of them are deployed without a security framework adequate to what they actually are.Securing the Thinking Machine is the practitioner’s guide to AI security for the people responsible for getting it right: security leaders building programmes from the ground up, engineers implementing controls, architects designing AI infrastructure, and governance professionals managing risk in organisations where AI is already operational.Written by AI governance practitioner Ian Loe, this book maps the complete attack surface of a modern AI deployment, from training data and model weights through inference engines, orchestration layers, memory systems, and autonomous agents. It explains why the deterministic security thinking that built the last generation of defences is insufficient for probabilistic systems, and what must replace it.What this book covers:The full AI stack as an attack surface, with threats mapped at every layer. The probabilistic failure modes that traditional security frameworks miss, including hallucination, sycophancy, goal misgeneralisation, and specification gaming. Technical defences applied layer by layer, from model supply chain security and inference engine hardening through prompt injection defence, guardrail architecture, and output filtering. The workforce security model for governing AI agents as autonomous actors with delegated authority. The FRAME governance framework and the AI Agent Evaluation Framework (AAEF), two original frameworks developed through direct enterprise experience. A five-level AI security maturity model for building and measuring a security practice. Regulatory mapping across the EU AI Act, NIST AI RMF, ISO 42001, and sector-specific standards. An assessment of the emerging frontier, including world models, embodied AI, persistent autonomous agents, and systemic risk at scale.This is not a theoretical survey of AI safety research. It is a working guide for practitioners who are today being asked to secure AI systems they may not yet fully understand, using frameworks built for a world that no longer fully applies.Who this book is for:Chief Information Security Officers and security leaders building AI security programmes. Security architects and engineers designing and implementing AI infrastructure. AI engineers and ML practitioners who need to understand the security implications of the systems they build. Governance, risk, and compliance professionals managing AI risk registers and regulatory obligations. Technology executives and board advisors responsible for AI strategy and oversight.