Episode #617
November 28, 2020
Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA) checks to see if you are human.
CAPTCHA is a kind of reverse Turing test.
The Turing test used two humans and a computer. One of the humans asked questions of the computer and the other human. If the questioner could not tell whether an answer came from the computer or the human more than 50% of the time, the computer was said to have artificial intelligence.
The CAPTCHA test is a computer checking to see if the user is a computer or a human.
CAPTCHA can ask a person to check a box if they are not a robot. The person may be asked to identify distorted letters in a graphic.
Another test is to check off boxes in a graphic grid that contain an object (like a street sign) … something a robot cannot do.
An Internet robot (BOT) reads text via a code (ASCII) which represents letters, numbers and symbols. “iMentor” for example, is coded: 105 77 101 110 116 111 114.
A graphic is coded by colored dots. Performing shape recognition from a pattern of dots is beyond the abilities of a BOT, so it cannot pick out objects from within a graphic. Only advanced, powerful programs and humans can identify objects within a graphic.
Attacks from BOTS are the reason for CAPTCHA on most websites. 
This iMentor Hints and Rants bLog website was inundated with spam comments until i added a CAPTCHA challenge. Most of the comments were Russian.
Many comments were disinformation regarding the recent Presidential election. It was common to receive 10 to 15 spam comment requests per day before adding CAPTCHA. Now i receive 2-3 spam comments per week, manually entered by humans.
i must approve all comments anyway, but having CAPTCHA to run interference makes the task much less time consuming and much less annoying.
The type of CAPTCHA i use on the bLog is free from Google. It is called reCAPTCHA.
reCAPTCHA is an invisible CAPTCHA. If a human fills out the comment form reCAPTCHA can tell from the typing pattern that he/she is human and allows the comment to be submitted. If an automated attack is suspected a graphic CAPTCHA is presented for the bot to try and solve.
Attempts to log into my WordPress admin page are subject to reCAPTCHA as well. Some identity thieves will commandeer a website to insert a fake page. You have probably received email that pretends to be a trusted institution, like Apple. If you follow the link to “confirm your account” (and i don’t recommend doing so) you will see that the fake Apple page is located on someone else’s website. These victims don’t use CAPTCHA to protect their admin.
i know that reCAPTCHA is working on my website because i am occasionally challenged when trying to log into my own admin page.
While reCAPTCHA’s algorithm is really good, bot makers are already working on the challenge. On goes the battle between the good guys and the bad guys.
We the humans (the good guys) are winning for now.
Fascinating!
Great information
I read somewhere that in addition to detecting whether the respondent is human or not, Google uses the picture CAPTCHA to add to its ever-growing database of image recognition. For example, if the CAPTCHA says click on all the images that contain a traffic light, Google uses the boxes that you click on to help it determine what a traffic light looks like, and it stores that information. Might just be Big Brother paranoia, but it makes a lot of sense.
Very smart of Google, Tom. But then smart and Google are redundant.