We Need to Talk About the Black Boxes

Last week saw two fascinating (and somewhat contradictory) announcements from Facebook. The first is that they will now work on code to detect and flag fake news, after months of claiming that this was a non-issue. The second is the news that Facebook are looking to finally enter China by implementing tools to allow third parties to censor content.

Behind these announcements (and many others recently from the world of Google, Twitter, and AI) lies a disturbing trend. Our world is increasingly being shaped by black boxes. These black boxes contain code that we cannot see. This code influences our news, entertainment, and relationships in ways that we’ll never understand.

In this sense, it’s never been more important to understand code. And over the past few years that importance has been met with dozens of “Learn to Code” manifestos and programs (even Obama was on board). As we head into a new year, there will no doubt be calls for another “Year of Code”.

But that’s not going to solve the problem. The world doesn’t need to be full of coders any more than it needs to be full of dairy farmers or mechanics or economists. The number of people in the world who need to actually write code is very small. But the number of people in the world who need to understand how code affects their lives is very large. The challenge is not to teach people to “Hello World”, it is to explain how code affects us every day. And that is a much harder challenge.

There’s one thing missing from all the cheery “Learn to Code!” websites. Learning to code, properly learning to code to the point of actually understanding how powerful code can be, is actually quite hard. Yes, printing “Hello World” is relatively easy in most languages (most). But the distance between that and understanding how code shapes our world is immense.

Fully learning a technology in order to understand how it shapes our world is a precedent that seems to be unique to computer programming. To take our dairy farmer as an example - every person drinking milk hasn’t gone off and spent a week on a farm understanding the entire process of grass –> cow –> milk. Let alone the packaging, logistics, and sales machines that gets that milk to your fridge.

Similarly the average person doesn’t spend years studying economics. But they do understand how interest rates affect their mortgage, and that if China slows down then Australia won’t have as much money to build roads and hospitals and schools.

The number of people who actually make a living from writing code is relatively small (and in the US, actually declining). So teaching people to code doesn’t really make sense in this respect either.

So why has this push been so strong? Because people who do code understand the power they possess. And they want more people to be aware of that power. The push came from the right place, but has been executed in the wrong way. In some ways it’s even backfired, creating and amplifying a division between people who can code and people who can’t. People who are in control and people who aren’t. Every person who has started on their “year of code” only to drop off a week later has this view of the world reinforced.

The black boxes are only going to become more numerous and more complex. Every change to a Facebook or Google algorithm makes that black box more complex. Every new AI advance is another box we can’t open.

It’s more critical than ever that people understand how code shapes their world. That won’t happen by teaching everyone to code, but through giving everyone the language to ask the right questions about code.

- November 2016