I’m not going to lie, I put this book off for a while because my kindle told me it was a 5 1/2 hour read. That shouldn’t have put me off because I really thought it looked like an interesting book, and it turns out I was mistaken anyway! The book finished at 66% through and the rest of the book was taken up by references from all the research that the author had cited.
This book takes us into the hidden world of algorithms. We’ve all heard of them and they sound great, futuristic and fancy. But as this book proves, algorithms are not perfect, nowhere near it. And it’s way too easy to imbue your algorithm with either your own or historical biases that make the algorithm inherently flawed and never fit for purpose.
Fry gives many examples of how algorithms have been used, from seemingly innocuous reasons like playing chess against grandmasters, and learning how to play jeopardy, to more serious applications like cancer diagnosis and deciding whether prisoners should get bail. We are taken behind the scenes to see how these algorithms came about, and what the problems can be when the algorithms are not considered carefully.
And our relationship with algorithms is interesting too – Fry talks about the fact that we have a tendency to over-trust things that we don’t understand, until we realise that it can make a mistake. And as soon as we know that an algorithm can make a mistake, we will then revert back to our own knowledge, even if our own knowledge is more flawed than the algorithm. So intriguing.
I was completely hooked from page one of this book, I read it all in one sitting as the writing style was so friendly that even though some of the topics covered were complex, I never felt confused about what I was reading.
For anyone confused about what an algorithm even is, or for those who are concerned about the increasing prevalence of them in our society, I’d definitely recommend this book. It’s a few years old now, but I think the points raised are just as relevant (perhaps even more so) as when they were written.
Leave a Reply