Have you ever been on a website and hit one of these:
What is it and why do I have to tick the box? Well, it’s called a “Captcha” and it’s a way for a website to protect itself from getting spammed.
As an example, a lot of people I know have been voting recently for the Kraft Hockeyville competition (Go Tata!) via a website voting page. Now if there wasnt a Captcha on the page, it’d be possible to write a program to just repeatedly click the page and submit thousands of votes. By placing a puzzle on the page that only humans can understand, you can stop this from happening.
But did you know you’re actually teaching a computer in the process?
Captchas initially looked like this (you might have seen these before):
This all stems from an attempt to scan 1000s of old books that are out of copyright and make them searchable online.
To do this, the books are scanned and then “read” by computers using a technique called “OCR” (Optical Character Recognition). But computers aren’t as good as humans at this job. In the example above, a computer would be able to make out the “$9000” but would probably struggle on the “starker”.
So to help out, each time you complete one of these captchas the service is using you to help read the text. It puts two words up, one of which it knows, the other it doesn’t. It uses the word it knows as a test to verify that you’re not cheating and writing random words.
Recently the system has been extended to include images – including those taken for street view maps. You have to wonder what benefit this is to Google – is it to help their self driving cars read roadsigns?
So next time you hit one of these tests, remember that you’re helping a computer learn to read text or maybe a car learn how to read road signs!