I can’t prove I’m a human

It often happens that I try to sign up for a website or service, and end up with something like this:

Supported browser? Why, I upgraded just yesterday!

Or sometimes, it’s even worse:

Again? But I haven’t even got to do it once!

What isn’t working is, of course, the reCAPTCHA service. That’s Google’s version of the CAPTCHA, or Completely Automated Public Turing test to tell Computers and Humans Apart. The purpose of a CAPTCHA is to prevent spam by only allowing humans through, and blocking automated machines.

In the beginning, the way CAPTCHAs worked was to give you some squiggly words, something like this.

Image credit: Peter Lindberg on Flickr

You were asked to type in these words into a box, to prove you can read them. Humans could read these words easily (well, sometimes not-so-easily), but machines couldn’t read them at all. So if the words were typed in correctly, it prove you were a human, and let you join the network, or post the comment, or whatever it was you wanted to do.

And so we had a lot of people spending time reading squiggly bits of text and copying them down.

That was when computer scientist Louis von Ahn, from Guatelama, came up with an idea to turn all that effort into something more useful. Having been one of the early CAPTCHA developers, he realised he had, to quote The Walrus, “unwittingly created a system that was frittering away, in ten-second increments, millions of hours of a most precious resource: human brain cycles”.

That was how reCAPTCHA was born. To the user, it looked just like a normal CAPTCHA, except that it had two words to type in, instead of a random string of letters. The difference was that, as they typed in the words, they were also helping to turn scanned paper documents into digital text.

Here’s how it worked: the user would be shown two words. The computer knew what one of the words was, but it didn’t know the other one. (But the order was random, so you could never tell which one the computer knew).

Image credit: combination of images by Scooooly and by BMaurer, from Wikimedia Commons

Both these words were taken from scanned books or other documents. If you got one word right, the computer assumed you got the other one right, too. And, over time it would “learn” what the new words were. So eventually, all the scanned books it had could be converted to digital text for people to read.

That’s why, if you look carefully, you’ll notice that reCAPTCHA came with the tagline “stop spam, read books”.

reCAPTCHA was a big success, and was later acquired by Google in 2009. And then, in 2013, Google announced reCAPTCHA version 2, also known as noCAPTCHA. That was partly because machines were getting smarter at reading CAPTCHAs, so the humans had to get smarter as well. And it was getting to the point where things were too hard even for humans to read.

The new noCAPTCHA is much simpler. All you get is a checkbox to click on.

Image credit: Me :)

That’s because noCAPTCHA uses behavioral analysis. So it looks at the subtle differences in the way you click, tap, or move the mouse, to figure out if you’re a machine or a human. If (and only if) it can’t make out, then it gives you a harder challenge to prove yourself.

Image credit: WIRED

But there’s one catch with Google’s new CAPTCHA system: it needs more bandwidth. And I have a slow connection.

The new reCAPTCHA has no fallback option, even for people with older browsers or slower Internet connections. The result is that people like me can’t get past them at all — and, as a result, I can’t access the websites and services that use them, either.

Many people have criticised reCAPTCHA, saying that it’s especially hard for people with disabilities to get past. What they don’t mention is that for noCAPTCHA, the disabilities aren’t always physical. They can also be digital.

Don’t get me wrong. The new noCAPTCHA is usually a step forward. When it works, it’s much more user-friendly than the old ones, and helps train image recognition machines in the process. But unless and until they implement a graceful fallback for slower connections, I’ll have no choice but to keep on reloading the page and hope it eventually works. Which kind of defeats the purpose, because it’ll turn me into a machine….

I eventually managed what I set out to do, thanks to a new data pack which gives good initial download speeds. Now excuse me as I go and finish all the CAPTCHAs, while I can 😉



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store