A deeper Look into CAPTACHA Systems

Original article was published by Prachi Gopalani on Artificial Intelligence on Medium

A deeper Look into CAPTACHA Systems

Source: nakedsecurity.sophos.com

Those who are not familiar with “CAPTACHA” these are weird little quizzes you end up taking while landing on a website where you have to type-in jumbled sequence of words, numbers and letters to prove that you are not a bot.

We’ll discuss here history behind they came to be and the problem they are gonna solve, what implication it has and ethics involved.

What CAPTACHA stands for?

An acronym for “completely automated public Turing test to tell computers and humans apart” coined in 2000 which tests whether you are human or not. So, Turing Test was named after Alan Turing who was computer scientist mathematician which came up with question Can a computer talk like a human being? A game was conducted where a human judge didn’t know whether the answer given is from human or computer. So, we can say CAPTCHA is form of Turing test to make sure someone is human or not.

But the question is Why CAPTACHA?

As, you might have seen on many sites there is always someone or kind of bot saying “You won $10,000” or other kinds of problematic statements, so to cut these kinds of statement/spams at first place CAPTCHA was introduced. This was a barrier for bot threats just by taking simple human test/Turing test.

The more we label these kinds of data the smarter the AI became.

How can you beat a system where AI is becoming better? As, we all know ability of machine learning algorithms to learn text, objects within images, increasingly smart artificial intelligence (AI) is the reason why these challenges are becoming tougher and tougher.

Disabled users?

But what about all those people who don’t have ‘normal’ abilities? These are group of people who relies on AI to do this work for them, suppose an CAPTCHA says identify all the images of “car”, click on those then you can move to the website, if you are vision-impaired you can’t actually do that, then you need an image-recognition tool to help you by clicking on the “car images” and letting you through the website, but we can say if CAPTACHA is trying to avoid that kind of AI then you are blocking an person who is not an span bot but an actual human being which relies on AI.

Another question arises, how can you determine whether it’s a spam bot or human using AI (because of its disability).

Advancement of Machine Learning

There is broad history of CAPTCHA knowing that there is lot of different developments in-place, which says that image recognition or word/text recognition can’t be understood by computer or AI cannot read it. But with the growing sophistication of artificial intelligence and machine learning, any test can be used to train software (in this case, bots) to solve the challenge. In fact, this isn’t a recent advance. Machine learning algorithms began getting better at solving CAPTCHAs than humans years ago.

But as AI develops and gets better in image and text recognition it can outbreak the CAPTACHA systems. So, how do you build a system that can reasonably distinguish between who’s AI and who’s AI being powered by human?

Because right now its biased to the people who use AI to navigate the world, thus this means we have to design a new framework? May be with the technique of tracking system we can know mouse moving in the way human spontaneity. Which can determine human or not? Now we have data also how human works and moves mouse and type on keyboard.


Companies like Google, working to advance the CAPTCHA services, and working on next-generation having broader analysis, called reCAPTCHA v3, use a mix of cookies, browser attributes, traffic patterns, and other factors to evaluate ‘normal’ human behavior.

Now there are some Ethical concerns in general around CAPTCHA

(1) reCAPTCHA using human powered to translate these documents that’s good because we are digitizing and using data of how human behaves on a computer.

(2) And may be data being analyzed may or may not be sensitive it has to do what you being doing on internet.


For CAPTCHA there are different issues you can say to just redesign it because there are multiple aspects that need to redesign. CAPATCHA is going to be the dilemma as we develop more technology in the future. It’s important to recognize that there is more ethical consideration of CAPTACHA and issues for disabled folks is very real and very pertinent. And do I really want CAPTACHA system tracking how I use internet and judge whether I am human.

“CAPTACHA will becomes outdated as AI grows”