Just Finished Reading: Weapons of Math Destruction – How Big Data Increases Inequality and Threatens Democracy by Cathy O’Neil (FP: 2016) [218pp] 

Algorithms – it seems to be THE word of technological age we’re living through. They seem (and indeed are) everywhere. Most people see them plainly on Amazon recommendations and other seemingly innocuous, and to be honest sometimes hilariously bad ‘suggestions’ from other websites. But like the vanishing icebergs most of the work of algorithms happen DEEP in the background. Been turned down for bank loan, mortgage or online job application? Watched a convicted criminal get an eyebrow rising super low or super harsh sentence? Heard about ‘blips’ on the Stock market that suddenly buy or sell vast amounts of stock and then almost instantly buy it back again? Started receiving some oddly specific news on your Facebook feed? That’s probably an algorithm (or a few of them) at work doing what they do – making assumptions, making deals and changing (and sometimes destroying) people’s lives based on data, correct or otherwise. The author calls this sort of thing ‘Weapons of Math Destruction’ or WMDs for short. 

Of course, the problem with using math (or as we say over here, maths) to make decisions about people's lives or livelihoods is that human beings are messy. Few of us (fortunately) fit into neat categories and a great deal of what we do, know, believe and act on isn’t exactly easy to quantify. So, what’s a mathematician or statistician to do? Use proxies – in other words something that can (at least theoretically) be measured for something that can’t. Like the chance of defaulting on a loan – how do you measure something like that? Well, you compare the data of the person asking for money with people like him – so where they live, education, criminal record, criminal records of their neighbours, family and friends. In other words, you ‘profile’ them. Rather than treating them as an individual you treat them as a member of a group – be in women, the young, racial minorities and so on. It makes decision making by the bank SO much easier. But if you’re an upstanding  member of a group with a bad rep? - Sucks to be you, right? And in a nutshell, there’s your problem. Unlucky enough to be born in a ‘bad’ area or into a risky demographic, you’re screwed. Born on the right side of the tracks? Your life is going to be a lot easier – especially with a fair winded algorithm at your back. 

Using 99.9% US examples – apart from 2 UK examples (one of which was hilariously from the British ‘city’ of Kent) - covering the whole gamut from school grades, teacher standards, incarceration suggestions, loan applications, and much more this was an often-frightening look at how pervasive algorithms are and how much effect they have on millions of people's lives – often with unthinking and unwarranted approval. Despite being almost entirely US focused and starting to feel rather out of date even after just 7 years, this will be a wake-up call to anyone not already familiar with the implications of algorithm influenced decision making. Many more questions need to be asked about how they work, what assumptions are deeply embedded within them and how we go about correcting badly performing ones. A recommended read (with some caveats) to anyone who has wondered what all the fuss is about.

[Labels Added: 1, Labels Total: 58]