Saturday, August 13, 2016

Weapons of Math Destruction | Catherine O'Neil

*Image and book provided by NetGalley in exchange for an honest review.


Today's world is run by algorithms. From our credit reports and finances to how well teachers and employees are doing, algorithms guide business in their decisions. They praised as being non-biased since they are programs, not people. Yet, as O'Neil shows us, the fact that they are made by humans and built upon data from prejudiced humans, the algorithms are not so fair as they are made out to be.


I first found out about this book because I follow O'Neil's blog, Math Babe. So when it came up on NetGalley, I immediately requested a copy. As someone who reads her blog, little of the information presented was surprising, but the depth that she was able to add in this book was very helpful.

The first several sections of the book talk about the author's relationship with algorithms. She worked in the finance industry and was only given so much information about each "black box" algorithm. This way no one employee could have all the secrets to what made the algorithm work. Unfortunately, this lack of transparency ended up hurting a lot of people. And it wasn't just the financial industry, either. Many companies, as well as the government, use algorithms based on data that may be incredibly irrelevant to whether or not someone will be a good employee or should stay in jail longer or shorter times, or should pay higher insurance premiums. Many aspects of humanity can't be quantified, so the algorithms attempt to fill the void by using other data. The problem with that is that there is no feedback on whether or not it works so the algorithm cannot adjust itself.

While the majority of the book is explaining these algorithms and how they hurt people, I was very grateful that O'Neil included a section on proposed solutions. I was worried the author would only complain about the problem throughout the book, but there were ideas for making things better. More importantly, to me, every claim she wrote about why the algorithms are broken or need transparency had a reference to back it up. O'Neil presents the problem, proposes solutions, then enables the reader to follow up on her research.

This is a good book for anyone looking to see how mathematics can help and hinder real life. This is also a good book for anyone interested in learning more about Big Business. Heck, even if you just need an explanation for why you pay so much for car insurance, give this book a read. 4 hoots!


                Hoot! Hoot!

No comments:

Post a Comment