This book puts into words many of the things that I’ve been feeling as an artificial intelligence skeptic. It explains the danger of trusting important decisions to computer systems that are black boxes, un-understandable even by their creators; highlights the ways that AI further concentrates power into the hands of relatively few; and even touches on how AI can strip our lives of what makes them meaningful.
On top of that, Chapman points out an important fact: AI’s primary use case is powering advertisement recommendation engines. It isn’t doing so well in other positive-impact areas like medicine, not to mention the negative impacts it’s having in crime suspect identification, loan approval, and other applications.
I didn’t love the presentation, though. The book feels a little fearmonger-ish, in a way that probably makes dismissing it easy for AI’s proponents. Many sections also seem underdeveloped, like Chapman is trying to cover too much in too little space; I would have enjoyed learning a bit more about how traditional science and engineering techniques can be used to keep AI in check and about ways we can speed scientific progress without AI. Maybe those are separate topics for separate books?
I was also a bit disappointed by the fact that Chapman doesn’t even attempt to encourage people to ditch advertising-funded, AI-driven services like social media. He encourages readers to take basic steps to limit online surveillance but fails to address the fact that no amount of ad blocking is going to stop Facebook from tracking your every interaction within its apps. The book seems to imply that people depend on Internet services and can’t be expected to give them up, but if AI recommender engines are as harmful as Chapman makes them out to be, then I think we ought to at least abandon the most egregious offenders.
Despite its flaws, the ideas in Better Without AI are good, and I hope that this book and others like it will succeed in pushing back against the current AI frenzy.
(I found this book thanks to Matthew Butterick’s excellent piece “Will AI obliterate the rule of law?”. It’s a brief essay that makes a good case for the fact that AI is being used to circumvent the law; I’d definitely recommend it.)