Trivia: How reCAPTCHA Works

By Daniel Scocco

I’ll start with some basic facts which you may already know, but at the bottom of the article there’s a curious fact that you might have not heard before.

First things first, what is a CAPTCHA? It’s an acronum for “Completely Automated Public Turing test to tell Computers and Humans Apart.” Now in simple words: it’s basically a test used to make sure that whoever is performing an action is a human.

Why would you want to verify that? To avoid having bots or automated scripts doing things inside your website, application, and so on, which could lead to security problems as well as an overload of the system.

There are several ways to implement the test, but the most used one is presenting the user a sequence of letters or words and asking him to re-type and submit those letters. The image below illustrates how it works:

repatcha

The image below is the CAPTCHA test from a company called reCAPTCHA, which was originally a project at the Carnegie Mellon University and later got acquired by Google.

Now here comes the interesting trivia: on all reCAPTCHA forms (so those on all Google services) you actually just need to type the first word to pass the test and prove that you are human. The second word is one that Google’s computers could not identify on its book digitalization project, and by typing it you are helping Google to identify those words and complete their book scannings.

Neat huh?

It seems to be working as well, as every day over 200 million forms like the above one are displayed around the web.



Share

6 Responses to “Trivia: How reCAPTCHA Works”

  • Madhav Tripathi

    Oh my god, this is the reason sometimes wrong captcha works. But I never noticed wrong captcha works only for second word.

  • Extreme John

    Brilliant! That could really help do more or less half the job. That’s a very interesting project. Captchas are one of the best ways to avoid spam nowadays especially that spambots are performing even better than before.

  • Matej

    …and I always wondered why I manage to type correctly words on Adwords Keyword Tool, even when I’m sure I couldn’t guess it all right, lol 🙂

  • Daniel

    Interesting.

    Matej,

    I think the Google Adwords Keyword tool has been set up that way. Basically I think I could just about type anything into the captcha and I would be able to do my Keyword research.

    I can assure you there have been so many times where the words were simply unreadable(total mess)so I just put in what I felt may be close.

    On the other side of the coin, I have put in the exact words required over at a pinging site(The catchpa is crystal clear compared to the Adwords one) yet, over and over again my entries were rejected. Eventually, I just went to another pinging site.

    And to Daniel(The Author) yes, I have noticed those second words(Often the clearer and easier to read of the two presented) and that is quite an interesting snippet of trivia.

  • Jerrick

    I still prefer using mathematics question and request to answer and verify as human.
    For Recaptcha, i still able find some robot spam comment that able to ignore those recaptcha and spam comment all over the blogs.

  • Sydney @ Social Dynamics

    I’ll try that trick the next time I encounter a Captcha feature. Though seriously, I hope they fixed their fonts, it’s so difficult to decipher sometimes that I wonder if even they, themselves could figure it out.

Comments are closed.