paint-brush
What is CAPTCHA and Does Google Use it to Train AI?by@Turner
3,575 reads
3,575 reads

What is CAPTCHA and Does Google Use it to Train AI?

by MatthewOctober 1st, 2021
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

CAPTCHA was invented between 1997 and 2000. It was used to protect against spam requests received every day in the now-defunct AltaVista search engine. CAPTcha offers a test that a person can easily solve, but a machine will not recognize. It helps various programs and services to identify robots and block their actions. There are many CAPTCHAs, here are the most common and frequently used CAPTchAs. The reliability of the CAPCHAs from the technical side may be questionable, but the program is undoubtedly helpful.

Companies Mentioned

Mention Thumbnail
Mention Thumbnail
featured image - What is CAPTCHA and Does Google Use it to Train AI?
Matthew HackerNoon profile picture

Officially, CAPTCHA protects the site from bots, spam, flood, and other evil spirits. The reliability of the CAPTCHA from the technical side may be questionable, but the program is undoubtedly helpful from a human and philosophical point of view.


However, isn't CAPTCHA just an excuse for certain users to train artificial intelligence?


When did the CAPTCHA appear?


CAPTCHA was invented between 1997 and 2000. It was used to protect against spam requests received every day in the now-defunct AltaVista search engine.


The term was officially introduced by a team of developers from Carnegie Mellon University in 2000. At the same time, users first saw annoying pop-ups asking them to confirm that they were not robots.


The CAPTCHA offers a test that a person can easily solve, but a machine will not recognize. In a basic CAPTCHA, a puzzle appears on the screen in the form of an encrypted text that needs to be entered in the appropriate field. Or it can be something more sophisticated.


CAPTCHA is indispensable for companies — especially for those whose activities are related to financial transactions. For example, when airlines open ticket sales, the technology cuts off bots that would instantly buy up all seats.


reCAPTCHA is a continuation of the CAPTCHA project, also developed at Carnegie Mellon University. The project itself was purchased in 2009 by Google and is supported and developed by its forces.


At the moment, this is the most optimal way to implement this test. The user does not even see that any protection is installed on the site for the most part. Therefore, it means that he does not experience any inconvenience.


Why do we need a CAPTCHA?


CAPTCHA serves as a barrier for robots. It helps various programs and services to identify robots and block their actions.


• Hacking protection


If there were no CAPTCHA, it would gradually be possible to hack a lot of accounts without two-factor authentication. That is, you created an account and protected it with only one password. How can this account be hacked? By constantly trying different passwords. The robot can generate thousands of passwords in a short period, and it is likely that sooner or later, it will be able to find the right one if this robot is not stopped with the help of a CAPTCHA.


• Spam protection


In general, there is a fierce struggle with spam on the internet. In addition to CAPTCHA, programs and services consider user complaints, analyze spam sources, and constantly try to invent new methods to combat malicious spammers.


• Overload protection


It's very easy to overload a website or service. It is done by sending them too many requests without time to work them out. These requests can just be sent using automatic malware.


A CAPTCHA is the saving remedy that will stop each of these threats.


Types of CAPTCHA


There are many CAPTCHAs. However, here are the most common and frequently used CAPTCHAs.


  1. Text recognition. A traditional type of CAPTCHA requires the user to enter a series of numbers and letters from an image. Most often, a text string is distorted by various colors and filters.
  2. Image selection. Such a CAPTCHA asks users to identify specific photos from a set (for example, specify all images with hydrants.)
  3. A logical task. It checks whether the specific user in front of it can think logically or not.
  4. Three-dimensional CAPTCHA. It requires the user to identify an image, letters, or numbers displayed in three dimensions. In principle, this is a complicated version of the previous types.
  5. Marketing CAPTCHA. It asks the user to enter a word or phrase corresponding to a certain brand.
  6. CAPTCHA "I'm not a robot." The veracity of this statement is determined by the time spent to execute the task and other more complex mechanisms. The CAPTCHA captures the time on your computer and the time zone. It determines the approximate location and the browser that is being used. It takes into account even the resolution of the screen on which it is displayed. Mouse movement and other means for cursor control and clicking are also monitored.


  1. Sound CAPTCHA. It presents the user with a series of pronounced letters or numbers. Often the user can request an alternative text display.
  2. CAPTCHA drag-n-drop. It can offer to assemble an image by moving its parts.
  3. Honeypot CAPTCHA. It places hidden fields on the screen that are invisible to humans but understandable to the bot. When the software starts interacting with the code, it becomes clear that the service is used not by a person but by a machine. Such actions will be blocked.


The CAPTCHA should reduce the likelihood of bots on the site. The service analyzes the risks and can often choose one of the above types of tests. But on a particular page, only one specific task can be used.


Why do most users dislike CAPTCHA?


Most online users are averse to CAPTCHA, even though it is a relatively convenient blocking bot.


  • CAPTCHA takes time. Each test takes only a couple of seconds. But if you count how many puzzles you have to solve in a week or a month, you get an impressive figure.
  • CAPTCHAs can be too complicated. Sometimes it can be difficult to distinguish between numbers and letters, especially if the characters are too distorted. For instance, you can confuse the letter "O" and the number "0."
  • CAPTCHA interferes with users with features. It is often too difficult for people who have vision problems. Some systems include the ability to voice characters, but this does not always help.
  • CAPTCHA spoils the service interface. Often, distorted fonts and various color solutions do not fit into the overall shape of the internet portal. That’s why the overall visual is disrupted.
  • CAPTCHA can even reduce conversions. At the sight of the test, users can leave the site to save their nerves. As a result, websites lose potential customers — the more complex the protection, the fewer visitors.


How to bypass a CAPTCHA


Netizens who prefer not to solve CAPTCHA use web browser extensions that allow them to bypass codes. Popular services include:


AntiCapture is a plugin for Google Chrome and Mozilla Firefox browsers that finds and solves a captcha on a web page.


CAPTCHA Be Gone is a utility that detects CAPTCHA on web pages, solves it and copies the result to the clipboard. It is suitable for Internet Explorer, Google Chrome, and Mozilla Firefox.


Don’t forget that bots become smarter through the efforts of specialists. So, CAPTCHA is also developing, and these programs may require updates from time to time.


Does Google use us through CAPTCHA?


Indeed many of us have heard that Google uses CAPTCHAs to train artificial intelligence. But, of course, this information has not been proven in any way.


Nevertheless, there are numerous facts that directly or indirectly confirm this. For example, a CAPTCHA appears too often, even without the apparent need for it. And what is more interesting, sometimes it even ignores user errors.


It's hard to imagine how much Google saves on specialists every day, using ordinary internet users instead to teach the image recognition system for free.


In addition, internet users can accurately read the blurred text on the facades of houses, thereby teaching artificial intelligence to find the correct house numbers on the map.


Or, for example, to arrange house numbers. If users do this through a CAPTCHA, then the Google specialists cannot place houses separately. Instead, the system will determine where each number is located and automatically put it on the map. After that, the user will recheck the houses' placement and mark the wrong numbers, thereby making edits to the map.


If this is the case, then, what — does Google make us work for them? It looks like it! Just imagine what contribution all of us together make to the great cause of teaching Google's artificial intelligence — for free.


Of course, a CAPTCHA is a noble idea that protects websites from spam and cyberattacks. However, the fact that the Google system uses the system for its own purposes casts doubt on the transparency of the CAPTCHA.