Looking for a suitable Christmas gift? A good book is always a great choice, even in electronic form. But what to choose? What one finds exciting can be a matter of taste, and it’s easy to miss the mark. So, how about something useful? Perhaps a book on Artificial Intelligence (AI)? Understanding what the AI hype is really about can be an investment in the future and a way to join the conversation. But please, nothing too dry or filled with complex mathematics. If you enter a bookstore or search online with these ideas, you’ll be overwhelmed by a wave of books.
Here are my five definitive recommendations: a fictional biography, a new type of atlas, a hands-on book for playing and experimenting, a book on the mathematics of AI that even non-mathematicians can understand, and a classic for those who want to know it all.
How It All Started: Maniac
Let’s start with context and history. The age of Artificial Intelligence didn’t begin with the release of ChatGPT. What connects AI with mathematical logic, military research, the atomic bomb, the Cold War, and game theory is less known. In his novel “Maniac,” Benjamín Labatut weaves these themes into a fictional biography of John von Neumann, the inventor of modern computer architecture. It’s a rollercoaster through recent history that ends with the duel between human and machine, the world’s best Korean Go player Lee Sedol, and the AI AlphaGo.
The Environment: Atlas of AI
For those interested in the impact of modern AI and not wanting to be dazzled by the promises of the tech industry, the second recommendation is essential. Kate Crawford’s “Atlas of AI” reveals that AI is more than a miraculous and intangible form of machine intelligence. In this book, we learn about the tangible, material side of the AI industry: from the world’s lithium mines to the click farms of the global south, where “ghost workers” prepare data for AI training for minimal wages, to automated workplaces, massive data archives, AI training camps, and the Pentagon’s algorithmic warfare team. According to Crawford, AI is primarily “a technology of extraction,” profiting from the exploitation of minerals, cheap labor, and an immense amount of data.
The Mathematics: Why Machines Learn
If you want to understand how this mysterious machine learning works, I highly recommend the third book on the list. Yes, it’s about mathematics, but it’s relatively simple. Anil Ananthaswamy repeatedly approaches the problem from different perspectives, making it possible to understand why and how machines learn.
Learning Through Play: Understanding Artificial Intelligence
What happens when an artist, a musician with a background in philosophy and software development, writes a book on the basics of AI? You get a book that’s anything but boring, despite the technical subject matter. The engaging texts and cartoons by Sophia Sanner are complemented by numerous example programs available online. You can try out how various methods work, how simple scripts piece together texts, and what language models do when you tweak their parameters.
The Textbook: AI, A Modern Approach
Your first encounter with this textbook by Peter Norvig and Stuart Russell might be a bit intimidating. It’s a hefty book with over 1,000 pages, covering much more than just machine learning and neural networks. When I first held this tome 15 years ago, I was surprised that Russell and Norvig started with a chapter on search algorithms—yes, search is AI too. The latest edition begins with a section on agents. Even if you’re not interested in all the technical details, you’ll find value in the history of AI and ethical questions. This book is the most cited textbook at universities for a reason.