I have started going to a local book club on science and tech books. This post was written to collect my thoughts in preparation for the meeting.

The book is about how humans should live with algorithms. In the authors own words:

It’s about asking if an algorithm is having a net benefit on society. About when you should trust a machine over your own judgement, and when you should resist the temptation to leave machines in control. It’s about breaking open the algorithms and finding their limits; and about looking hard at ourselves and finding our own. About separating the harm from the good and deciding what kind of world we want to live in. Because the future doesn’t just happen. We create it.

My main concern going into this book was that it would simply be a list of examples of algorithms doing wrong, like e.g. Weapons of Math Destruction. While reading one of this kind of books can be nice, reading a second would be a bit much. I wanted this book to suggested ways of solving the problems rather than add items to the list of problems.

The book didn’t disappoint, but I’m not overly excited about it either. I like the structure of the book – each chapter covers an area where algorithms influence our lives: law enforcement, healthcare, driving.

Things that I didn’t like

Each chapter contains a couple of examples oh how algorithms are applied in the area. unfortunately, the discussions were not as deep as I had hoped. Often they would cover 1-2 pages of a 30+ page chapter. The examples set the scene, but I didn’t get this book for the examples, I got it for the discussion that the example should facilitate.

What I liked the most

The concept of algorithm aversion was new to me. Hannah fry explains the concept like this:

But there’s a paradox in our relationship with machines. While we have a tendency to over-trust anything we don’t understand, as soon as we know an algorithm can make mistakes, we also have a rather annoying habit of over-reacting and dismissing it completely, reverting instead to our own flawed judgement. It’s known to researchers as algorithm aversion. People are less tolerant of an algorithm’s mistakes than of their own – even if their own mistakes are bigger. (…) This tendency of ours to view things in black and white – seeing algorithms as either omnipotent masters or a useless pile of junk – presents quite a problem in our high-tech age.

I’m sure it will be useful when we start putting algorithms to work. The doctors tell us that that they can tolorate quite high error margins – perhaps they have less aversion to algortihms than the rest of us.

There is one discussion point that I think was quite interesting and its the one about diversity and fairness in the prediction.

I liked the conclusion because it tackles the questions asked in the introduction. For example, the final paragraph of the book does a good job of describing her view on what future she would like to live in:

This is the future I’m hoping for. One where the arrogant, dictatorial algorithms that fill many of these pages are a thing of the past. One where we stop seeing machines as objective masters and start treating them as we would any other source of power. By questioning their decisions; scrutinizing their motives; acknowledging our emotions; demanding to know who stands to benefit; holding them accountable for their mistakes; and refusing to become complacent. I think this is the key to a future where the net overall effect of algorithms is a positive force for society.

Earlier in the same chapter she suggests that we can achieve some of these objectives if we start to build algortihms with the aim of making it easier to question their decisions. She has at least two suggestions for how this can be done: Make the algorithms give a confidence estimate in their predictions, and make them give several predictions at once.

I think this is something worth striving for, and I know that explainability and confidance estimates are very active research areas.

Questions for the book club

Do you feel better or worse about how algorithms incluence our lives after having read the book?
Did this book make you change your mind on any of the topics it covered?
Why aren’t we using Pigons more?
When you should trust a machine over your own judgement?
What are the machines limits? Whare are humanities limits?