captcha

Every captcha you've solved in the last 20 years you've been working for free. In 2007, a university professor noticed something interesting. That hundreds of millions of people are solving CAPTCHA every day, and all this brain power is spent on four crooked words. He said to himself, why not use this free labor? So he went on to build a new system called recaptcha. In this system, instead of one illegible word, it would show you two illegible words. The system knew one word, and that was actually the test. But the second word was scanned from a book or a real newspaper that even the computer itself couldn't read. If you type the first word correctly, the system would say, well, you typed the second and your answer will be recorded. I mean, you thought you were proving that you were a human, but you were actually transcribing books for Google. By 2011, the same system had digitized Books archive and 13 million articles from the New York Times. The story doesn't end there. Since 2012, recaptcha has changed. It didn't show the words, it showed the images. Every time he told you to find the sidewalk lane or select the stop sign, you were tagging the instructional data for Google's AI vision models, the same models behind Google Maps and, most importantly, self-driving cars. The next time CAPTCHA tells you to prove that you are a human, you're actually showing how to be a human.