CAPTCHA, an acronym for “Completely Automated Public Turing test to tell Computers and Humans Apart”, is a type of challenge-response system primarily used to differentiate human users from automated bots on the internet. It is a test that uses distorted visual or audio cues that are easy for humans but difficult for AI to solve.
Think of CAPTCHA like a doorman at a private event who asks everyone a secret question before they enter. If you’re a human, the question is supposed to be easy (like reading a squiggly word or selecting pictures of buses). But if you’re a robot, it’s often too hard because robots aren’t good at understanding distorted text or images like humans are.
CAPTCHA is a standard security technology that’s largely implemented to protect a range of areas online. These protected areas include but are not limited to user registration, online polling, e-commerce transactions, and password reset systems. This preventative measure works by introducing a verification step into the user authentication workflow. This step typically includes a test that is straightforward for humans to pass but designed to be difficult for bots, programs, or algorithms.
Its primary goal is mitigating the potential impact of automated cyber attacks, such as scraping, spamming, bruteforce attacks, and dictionary attacks by making it difficult for bots to impersonate humans. For CAPTCHA to be most effective, it must pose a problem that is easy for humans but difficult for algorithms to solve. For example, distorted text, image recognition tasks, simple mathematical questions, or audio transcriptions.
The distorted text CAPTCHA, the most common variant, frequently presents the user with distorted text input in an image and requires the user to accurately interpret and input this text in a response form. The distortion of this text presents a difficult problem for Optical Character Recognition (OCR) software, which AIs typically use to interpret written text in images.
However, there are different variants of CAPTCHA, such as image recognition CAPTCHAs, math CAPTCHAs, and even 3D CAPTCHAs. Despite variations, all CAPTCHAs share a common objective: their solutions should be easily discovered by humans but very challenging for AIs.