Guide
The best books to learn AI, machine learning and deep learning
Whether you want to understand AI for work or learn to build it yourself, there's a great book for your level. We've sorted these from no-maths introductions to the textbooks researchers use, and flagged the ones you can read free online.
Start here (no maths needed)
For anyone who wants to understand AI, use it well and see through the hype.
Best for using AI at work
1. Co-Intelligence: Living and Working with AI by Ethan Mollick (2024)
BeginnerA practical, upbeat guide to using AI as a co-worker, co-teacher and coach, from a Wharton professor who studies how AI changes work.
- No technical background needed
- Concrete advice you can use at work the same day
- Balanced on both opportunities and risks
Price: Paid bookFormats: Print, Ebook, Audiobook
Best for spotting hype
2. AI Snake Oil by Arvind Narayanan and Sayash Kapoor (2024)
BeginnerTwo Princeton computer scientists explain what AI can and cannot do, and how to spot claims that are too good to be true.
- Teaches you to tell real AI progress from hype
- Clear explanations of where AI predictions fail
- Useful for managers buying AI products
Price: Paid bookFormats: Print, Ebook
Best story of how AI learns
3. The Alignment Problem by Brian Christian (2020)
BeginnerA readable story of how machine learning systems learn values, why they sometimes learn the wrong ones, and what researchers are doing about it.
- Explains machine learning through real stories
- Covers bias, fairness and AI safety
- Well researched, with interviews with leading researchers
Price: Paid bookFormats: Print, Ebook, Audiobook
Most fun introduction
4. You Look Like a Thing and I Love You by Janelle Shane (2019)
BeginnerA funny, illustrated tour of how AI works and why it fails in weird ways, from a researcher famous for her AI experiments.
- The most entertaining introduction to how AI learns
- Great for teenagers and non-technical readers
- Makes AI mistakes easy to understand
Price: Paid bookFormats: Print, Ebook, Audiobook
Machine learning
For people ready to learn how machine learning works, with some maths or code.
Best short overview
5. The Hundred-Page Machine Learning Book by Andriy Burkov (2019)
IntermediateA compact overview of the main machine learning methods, with just enough maths to understand how each one works.
- Short enough to finish in a weekend
- Covers the essential algorithms and ideas
- Good refresher before interviews
Watch out: Assumes some comfort with maths notation.
Price: Paid book (read-first, buy-later option on the official site)Formats: Print, Ebook
Best hands-on guide
6. Hands-On Machine Learning with Scikit-Learn, Keras and TensorFlow by Aurélien Géron (2022)
IntermediateThe classic practical guide to building machine learning and neural network projects in Python, with code notebooks you can run.
- Learn by building real projects
- Free code notebooks on GitHub
- Covers both classic machine learning and deep learning
Watch out: You will need basic Python.
Price: Paid book; code notebooks are freeFormats: Print, Ebook
Best free textbook
7. An Introduction to Statistical Learning by Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani and Jonathan Taylor (2023)
IntermediateA widely used university textbook on the core ideas of machine learning, written to be accessible, with Python and R editions.
- Free to download from the official site
- Clear explanations used in courses worldwide
- Python edition with hands-on labs
Price: Free online; print edition availableFormats: Free PDF, Print
Best complete course in a book
8. Machine Learning with PyTorch and Scikit-Learn by Sebastian Raschka, Yuxi (Hayden) Liu and Vahid Mirjalili (2022)
IntermediateA thorough, code-first path from classic machine learning to deep learning with PyTorch, the most popular deep learning library.
- Explains both the ideas and the code
- Goes from basics to transformers
- Author writes the respected Ahead of AI newsletter
Watch out: Long; best treated as a course rather than a quick read.
Price: Paid bookFormats: Print, Ebook
Deep learning and building with AI
For going deeper into neural networks and building real AI applications.
Best practical start
9. Deep Learning for Coders with fastai and PyTorch by Jeremy Howard and Sylvain Gugger (2020)
IntermediateA top-down, hands-on book that gets you training working models first and explains the theory as you go. The basis of the free fast.ai course.
- Start building on day one
- Free to read as notebooks on GitHub
- Pairs with a free video course
Watch out: Uses the fastai library, which is less common in industry than plain PyTorch.
Price: Free online; print edition availableFormats: Free notebooks, Print
Best modern textbook
10. Understanding Deep Learning by Simon J. D. Prince (2023)
AdvancedA modern deep learning textbook with clear diagrams, covering transformers, diffusion models and more, with a free PDF.
- Up to date with modern architectures
- Excellent visual explanations
- Free PDF and practice notebooks
Watch out: University level; expect maths.
Price: Free PDF; print edition availableFormats: Free PDF, Print
The classic reference
11. Deep Learning by Ian Goodfellow, Yoshua Bengio and Aaron Courville (2016)
AdvancedThe classic textbook on deep learning theory by three of the field's leading researchers, free to read online.
- Deep, rigorous foundations
- Free to read on the official site
- Written by pioneers of the field
Watch out: Heavy on maths, and predates transformers and large language models.
Price: Free online; print edition availableFormats: Free online, Print
Best for understanding LLMs
12. Build a Large Language Model (From Scratch) by Sebastian Raschka (2024)
AdvancedWalks you step by step through coding a ChatGPT-style language model yourself, to understand how they really work.
- The clearest way to understand LLMs from the inside
- Every step in runnable code
- Covers pre-training and fine-tuning
Watch out: Needs solid Python and some PyTorch.
Price: Paid book; code is free on GitHubFormats: Print, Ebook
Best for building AI products
13. AI Engineering by Chip Huyen (2025)
IntermediateHow to build real applications on top of AI models: choosing models, prompting, evaluation, retrieval and agents. The most-read book on O'Reilly's platform in 2025.
- Practical guide to building AI products
- Strong on evaluation, often skipped elsewhere
- Useful for engineers and product managers
Price: Paid bookFormats: Print, Ebook
How we chose
We picked books that are clear, widely recommended by practitioners and still relevant today. We say which ones need maths or code, and we link to the official free editions where the authors offer them. Covers shown here are our own designs, not the publishers’ artwork.
New to the terms? Our glossary explains machine learning, neural networks, large language models and more in plain English.