REAL LIFE NEWS: CAPTCHAS MAKE IT MORE DIFFICULT TO PROVE YOUR HUMANITY
by Hazed
Have you noticed recently that CAPTCHAs - those annoying distorted bits of text we are supposed to identify and retype when setting up accounts on all kinds of systems on the net - are getting more and more difficult to decipher?
Several times recently I’ve found myself staring at the string of bendy, fuzzy letters completely unable to figure out what they are supposed to be, and having to hit the button to generate a new set multiple times before I can solve the puzzle.
The reason they are getting trickier is because of an arms race between hackers, who want their scripts to be able to solve them automatically, and the devisers of CAPTCHAs, who want to ensure only flesh-and-blood humans can use whatever service they are protecting.
CAPTCHA stands for “Completely Automated Public Turing test to tell Computers and Humans Apart.” They’ve been around for about ten years and are now ubiquitous. Which makes it more frustrating that they are becoming unusable!
Studies have shown that humans are finding the things harder to solve, and it has been reported that 1 in 5 visitors will give up and leave a website rather than attempting to complete one. As an example, Bella tells me she is one of the people who has stopped bothering with websites that use CAPTCHAs. She says she classes them with setting the video recorder to record programs, but adds, “Fortunately, we've now moved into a block of apartments where several families have small children, and I can usually borrow a five year old to do the programming for me...” Shame she can’t get them to do the CAPTSCHAs, too!
Meanwhile, machines have infinite patience and can keep on trying until they crack a CAPTCHA. Once they learn how to solve a particular type, the solution will get propagated around the net like lightning and every hacker will be able to write scripts to pretend to be human.
As the arms race continues, two things are likely to happen. The first is that hackers’ attempts to defeat more and more sophisticated protection techniques are going to start pushing the boundaries of artificial intelligence. They will have to teach their bots to think like humans.
The second thing is that humans will find it more challenging to solve the problems.
The result is that it may well be impossible for us poor, fleshy creatures to prove our humanity!
