A quick note before the list: I’ve been living in this space for a while now — as an instructional coach, a Google Certified Innovator, a doctoral student, and someone who uses AI tools daily in my actual work. The books I’m recommending here are ones I’d press into the hands of a thoughtful educator or a curious non-technical reader. This is not a developer’s reading list. If you want to build LLMs from scratch, you’re reading the wrong blog.
What I care about: understanding what these systems actually are, what they can and can’t do, what they mean for teaching and learning, and how to think clearly about the cultural and ethical questions they raise. The AI book market has exploded with hype, doom, and everything in between. Most of it isn’t worth your time. Here’s what is.
Where to Start
Co-Intelligence: Living and Working with AI — Ethan Mollick (2024)
This is the book I recommend first to every educator asking me where to begin, and it’s not particularly close. Mollick is a Wharton professor who has been using AI in his classroom since the day ChatGPT launched and writing about it — honestly and with genuine curiosity — at his Substack ever since. Unlike most AI books, this one was written by someone with actual daily practice rather than theoretical distance.
The central argument is in the title: AI as co-intelligence, not replacement intelligence. Mollick’s four rules for working with AI are practical enough to start using today and deep enough to keep thinking about. His concept of the “jagged frontier” — that AI is weirdly capable at things we’d consider hard and oddly bad at things we’d consider easy — is the single most useful mental model I’ve found for calibrating what to expect.
For educators specifically, Chapter 7 on AI in schools is worth the price of the book alone. Mollick is genuinely thoughtful about the implications for assessment, expertise development, and what we’re actually asking students to do when we assign traditional work in an era of capable AI tools. He doesn’t hand you easy answers. He asks better questions.
Worth noting: some readers already deep in this space find it a bit surface-level, and it was written in 2023, so some specifics are already dated. Read it for the framework, not the technical details.
Understanding What AI Actually Is
Artificial Intelligence: A Guide for Thinking Humans — Melanie Mitchell (2019)
Still the best accessible introduction to what AI fundamentally is and isn’t. Mitchell is a computational complexity researcher at the Santa Fe Institute, and she brings real intellectual rigor to a topic that attracts an unusual amount of noise. This book predates the LLM explosion, which is actually part of what makes it valuable — it gives you the conceptual foundation to understand why systems like GPT surprised even the researchers who built them.
Mitchell is especially good on the gap between narrow AI capability and what we loosely call “understanding.” If you want to have an informed opinion about whether AI is “really” thinking, read this first.
Get Artificial Intelligence: A Guide for Thinking Humans
The Coming Wave: Technology, Power, and the Twenty-First Century’s Greatest Dilemma — Mustafa Suleyman (2023)
This is the big-picture book. Suleyman co-founded DeepMind and Inflection AI before becoming CEO of Microsoft AI — he is, in other words, someone who has spent his career at the center of this thing. The Coming Wave is his argument that we are facing a genuine civilizational inflection point with AI (and synthetic biology), and that the window to build appropriate containment structures around these technologies is narrowing rapidly.
What distinguishes it from most AI doom-or-boom books is specificity. Suleyman doesn’t deal in vague anxieties — he makes concrete arguments about the concentration of power, economic disruption, and the structural problems of trying to regulate technology that spreads faster than governance can follow. Readable, serious, and useful for understanding why AI isn’t just a productivity story.
The Ethics and Alignment Problem
The Alignment Problem: Machine Learning and Human Values — Brian Christian (2020)
If you want to understand why making AI systems that reliably do what we want them to do is genuinely hard — technically, philosophically, and ethically — this is the book. Christian spent years interviewing researchers at the leading AI labs and built a rigorous, human-readable account of the problem at the center of AI safety.
The alignment problem isn’t abstract. It shows up in recommendation systems that optimize for engagement and produce radicalization. It shows up in hiring algorithms that encode historical discrimination. It shows up every time a system is optimized for a measurable proxy of what we actually want, rather than the thing itself. Christian is excellent on how this happens, why it’s hard to fix, and what the researchers working on it are actually doing.
This book complements Mollick’s more optimistic framing well. Read both.
Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence — Kate Crawford (2021)
The critical perspective this list needs. Crawford, a researcher at USC and co-founder of the AI Now Institute, makes a compelling argument that AI systems are not software abstractions — they are material, political, and economic objects with real costs and embedded power dynamics. The rare earths in the hardware, the data center energy consumption, the contract workers’ labeling training data in difficult conditions, and the labor displacement — Crawford maps all of it.
I don’t agree with everything in this book, and Crawford’s perspective is explicitly critical rather than balanced. But the questions she raises are important and underrepresented in the mainstream AI conversation. If you’ve read Mollick and want a counterweight, this is it.
The History and the People
Genius Makers: The Mavericks Who Brought AI to Google, Facebook, and the World — Cade Metz (2021)
The best narrative history of the deep learning revolution. Metz is a New York Times technology reporter who covers this beat obsessively, and he had remarkable access to the key figures: Geoffrey Hinton, Yann LeCun, Demis Hassabis, and the others who turned decades of dormant theory into the technology now reshaping every industry.
This is the book if you want to understand why everything changed so fast after 2012, what the competitive dynamics between labs looked like, and how the researchers themselves thought about what they were building. Reads like a thriller — the science is real, the rivalries are real, and the ethical stakes land harder when you know the people involved.
For Educators Specifically
Brave New Words: How AI Will Revolutionize Education (and Why That’s a Good Thing) — Salman Khan (2024)
Sal Khan founded Khan Academy. He’s also an optimist, which comes through clearly in this book. Brave New Words makes the case for AI as tutor, mentor, and educational equalizer — arguing that tools like Khanmigo can bring the one-on-one tutoring advantage (Bloom’s famous “two sigma” finding, that individual tutoring improves outcomes dramatically over classroom instruction) to every student who needs it.
I read this more critically than I read Mollick, because the institutional interests are more directly aligned with the argument. But the core vision — that AI could close genuine equity gaps in access to high-quality educational support — is worth taking seriously, and the specific examples from Khan Academy’s work are compelling. Read it alongside the Crawford book for balance.
The Short Version
If you read only one: Mollick’s Co-Intelligence. It’s the most practical and most directly relevant to anyone working in education or doing knowledge work of any kind.
If you want the big picture: Suleyman’s The Coming Wave. The most serious argument about what’s actually at stake.
If you want the history: Metz’s Genius Makers. The best story of how we got here.
If you want the ethics: Christian’s The Alignment Problem for the technical/philosophical dimension, Crawford’s Atlas of AI for the political/material dimension.
These books sit alongside my broader reading on technology and education — if you’re interested in that context, the Zettelkasten post covers the note-taking system I use to actually hold onto what I read across all of this.
The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!