Best books for AI Security Specialists

Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow

by Aurélien Géron · 2017

~860 pages~22 hrs to read

A practical tour of classic machine learning and deep learning with working code. Read it first if you come from security and need to know how models are trained and where they can be manipulated.

Machine Learning and Security: Protecting Systems with Data and Algorithms

by Clarence Chio, David Freeman · 2018

~390 pages~10 hrs to read

Shows how machine learning is applied to security problems such as malware and fraud, and how attackers can target those models. A good bridge between the two fields.

Hands-On Large Language Models: Language Understanding and Generation

by Jay Alammar, Maarten Grootendorst · 2024

~400 pages~10 hrs to read

Illustrated introduction to how large language models work, from tokens and embeddings to retrieval and fine-tuning. You need this mental model before attacking prompts, RAG pipelines or agents.

Not with a Bug, But with a Sticker: Attacks on Machine Learning Systems and What to Do About Them

by Ram Shankar Siva Kumar, Hyrum Anderson · 2023

A readable survey of real attacks on machine learning systems and how organisations can respond, written by AI red team practitioners. A solid orientation before the technical texts.

Adversarial Machine Learning

by Anthony D. Joseph, Blaine Nelson, Benjamin I. P. Rubinstein, J. D. Tygar · 2019

~340 pages~9 hrs to read

An academic treatment of attacks on learning systems and the defences against them, including poisoning and evasion. Read it for the formal threat models behind the practical tools.

Adversarial AI Attacks, Mitigations, and Defense Strategies: A Cybersecurity Professional's Guide to AI Attacks, Threat Modeling, and Securing AI with MLSecOps

by John Sotiropoulos · 2024

Covers attacks on AI systems, including generative AI, with threat modeling and mitigation practices for secure AI development. A practical companion to testing your own LLM applications.

Machine Learning for High-Risk Applications: Approaches to Responsible AI

by Patrick Hall, James Curtis, Parul Pandey · 2023

Covers governance, explainability, robustness and security for models deployed where failures are costly. Useful when you advise teams on safe deployment and writing evaluations.

Threat Modeling: Designing for Security

by Adam Shostack · 2014

~590 pages~15 hrs to read

The standard method for finding design-level security problems before code is written. Apply it to AI systems by mapping data flows through models, tools and retrieval sources.

Building Secure and Reliable Systems: Best Practices for Designing, Implementing, and Maintaining Systems

by Heather Adkins, Betsy Beyer, Paul Blankinship, Piotr Lewandowski, Ana Oprea, Adam Stubblefield · 2020

~560 pages~14 hrs to read

Google practitioners on designing, running and recovering systems with security and reliability as shared goals. Helpful context for fitting AI controls into a production environment.

Security Chaos Engineering: Sustaining Resilience in Software and Systems

by Kelly Shortridge, Aaron Rinehart · 2023

Argues for deliberately testing how systems behave under attack to build resilience instead of aiming for perfect prevention. Relevant to leaders planning red team programs and resilience exercises.

Human Compatible: Artificial Intelligence and the Problem of Control

by Stuart Russell · 2019

~340 pages~9 hrs to read

A non-technical argument about controlling increasingly capable AI systems. Gives senior practitioners vocabulary for safety discussions with executives and policymakers.

The Alignment Problem: Machine Learning and Human Values

by Brian Christian · 2020

~360 pages~9 hrs to read

Narrative history of how machine learning systems come to behave contrary to human intent, and the research on fixing it. Read it for context on safety evaluation work.