You've probably seen these tests as well - an image that has certain letters in it which you must type in to prove that you are a human being. That is, for as long as computers can't do a good job analyzing these images and discovering the letters. Unfortunately, the Captcha project reports that this has already happened:
Thayananthan, Stenger, Torr, and Cipolla of the Cambridge vision group have written a program that can achieve 93% correct recognition rate against ez-gimpy, and Malik and Mori have matched their accuracy. Their programs represent siginifcant advancements to the field of computer vision.But don't worry! The good folks at Captcha are hard at work on the next set of tests that will help separate the humans from the machines... Try out ESP-PIX for example, in which you have to evaluate four images and choose the best word to describe what the images have in common... Not only does it keep out the machines (for now) but it is also child-proof!
But there is an easier way to defeat these things. A number of people have already documented systems that parse out the problem of defeating a Captcha challenge to an eletronic sweatshop in India or China -- hundreds of people who go to work each day to sit there and defeat the Captcha challenges...
Why would anyone go to the trouble of having a room of people in India defeating Captcha? Unfortunately lots of reasons. The same reasons that companies like Google are now using Captcha on Blogger -- one example: if you can use computers to automatically generate web pages that link to your product, you can elevate the place that your prouduct shows up in results pages... ditto product rankings or the results of surveys... In factm Captcha got started because of an online poll about the best computer science school:
In November 1999, http://www.slashdot.com released an online poll asking which was the best graduate school in computer science (a dangerous question to ask over the web!). As is the case with most online polls, IP addresses of voters were recorded in order to prevent single users from voting more than once. However, students at Carnegie Mellon found a way to stuff the ballots using programs that voted for CMU thousands of times. CMU's score started growing rapidly. The next day, students at MIT wrote their own program and the poll became a contest between voting "bots". MIT finished with 21,156 votes, Carnegie Mellon with 21,032 and every other school with less than 1,000. Can the result of any online poll be trusted? Not unless the poll requires that only humans can vote.So, where will this all end? Biometric challenges. It has to happen. How else will we be able to tell the machines apart from the people?