I have received a free ARC of the book from the publisher via NetGalley. This did not influence my opinion.

I have fallen for it and it’s my own fault: I expected this book to be interesting. It’s an academic book. My first hint should have been the publisher, Cambridge University Press. However, when read without this bit of information, the blurb borders on… wait for it… manipulative (cringe face). ‘New York Times bestselling author’ and ‘lift[ing] the lid on […] manipulation by artificial intelligence, algorithms, and generative AI, as well as threats posed by deepfakes, social media, and ‘dark patterns’ made me think, yes, this is the book I want to read.

It has taken me two weeks to get through the 218 pages (including acknowledgements and the opening matter) because the book is dense. The author devotes a chapter to defining the word ‘manipulation’ in various context, and this chapter is basically philosophy. The most interesting parts of the book, for me, were the screenshots of dark patterns taken from Dark Patterns Hall of Shame (https://hallofshame.design/) which I could have easily found on the website itself.

In chapter 7, where Sunstein finally reaches generative AI’s manipulatory (or not) practices, large chunks are devoted to the prompts the author used and lengthy quotes from the replies ChatGPT generated. I didn’t find this to be either particularly useful or interesting, especially as ChatGPT is updated on nearly daily basis (see an article from April 2025 where ChatGPT praises the user for stopping their mental illness medication – this update was introduced and withdrawn over the course of two days). I nearly DNFed before getting to chapter 7 and I am glad I decided to finish reading the book, although I did skim parts of chapter 6 because, frankly, at some point the examination of words and their context became so abstract I was no longer sure of what I was reading.

Overall, I feel somewhat manipulated (sorry) by the blurb, although I admit I am guilty of doing most of the manipulation myself – here’s where the idea of nudges, introduced by the author, kicks in. I guess I got nudged – I read the blurb I wanted to read, not the one that was there; I missed the publisher’s name; and then I am here complaining about having received the book I should have expected, rather than the one I was expecting. If you’re searching for an academic deconstruction of what ‘manipulation’ means; how, where, and when it is used; what are its dangers and what happens when it isn’t used enough – this is the book for you. If you think you’re going to read a book explaining why AI is either good or bad for you, as I expected, this isn’t it.

My ratings:
5* = this book changed my life
4* = very good
3* = good
2* = I should have DNFed
1* = actively hostile towards the reader*