The Alignment Problem is just an excellent book about the state of AI philosophy and ethics at the beginning of the 2020s. Because it’s more about ethical and philosophical problems, than technological ones, it’s much more in my wheelhouse. But I think that these questions are extremely important. And Christian just does an excellent job of laying out all the aspects of these problems in a highly readable, engaging and occasionally funny way.
Though this book is about AI, Christian ranges far and wide in discussing the sources of thinking about the ethics and broader philosophy of AI. That is because the subjects of book have done so, but a lesser writer could have just taken these ideas on faith instead of going back to the earlier sources. (For example, he briefly detours into medieval catholic moral philosophy.) From the captivating introduction to the conclusion, Christian grounds these problems both in the historical tradition of ethics (and philosophy in general) and in the real-world considerations that should be governing how we think about evolving technologies.
It’s a masterful display of how to write about technology ethics and philosophy in a way that is both highly accessible and also thoughtful and thought-provoking. But it’s also just an entertaining read. And that’s one reason why I recommend this book so much. It’s a page-turner in a way that so many books about ethics are not. (Trust me, I know.) Having a guide who is passionate about the subject matter but who is also a great communicator is key to making important issues like these more mainstream. And Christian does that in just about every section of this book.
For me it’s a must read, at least until the state of ethical and philosophical issues surrounding AI change materially (which isn’t likely to happen soon). One of the best books about the ethics of technology I’ve ever read.