Did you know that Google and a handful of other smart companies are relying on your free work to enhance their services and make a profit? On this episode, I talk about how ReCAPTCHA (a Google company) has utilized your entries to digitize books, to enhance Google Maps, and most recently to train Google’s autonomous car algorithm to detect cars, street signs, and road hazards. If you’ve ever filled a contact form, you’ve probably worked for Google..
CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) stems back to 2000. You may remember how contact forms asked you to answer a simple math equation before submitting the form. Back then answering the simple math equation only proved that you are human that knows 1+3 = 4. It did not do anything awesome. Instead, it just seemed annoying.
In 2007 Professor Luis von Ahn and four of his college students from Carnegie Mellon University (correction from video) realized that they could be doing something with all of this free crowdsourced work that people were doing to prove that they are not bots. They replaced the math equation with two words, creating reCAPTCHA. The genius to their method lies in how the two words were generated and how the responses were used. Their goal was to digitize books, so the words that appeared were scanned from books (separated by an algorithm that detected whitespace and used it as an identifier to distinguish one word from another). One of the words that appeared was already defined by reCAPTCHA (the control word), so if you answered that word correctly, the system knew that you were not a bot. The second word, however, was not yet known and if you defined the control word correctly, it was assumed that you would also answer the unknown word correctly. When answering the second word, you were actually defining that word for reCAPTCHA. Once enough people answered that word the same way, it was marked as valid, helping to digitize books. Google took notice of this awesome process and acquired reCAPTCHA in 2009, later using it to enhance other Google products.
Shortly after selling reCAPTCHA to Google, Luis von Ahn started DuoLingo. DuoLingo was presented as a language-learning app on the front-end and sold translations to companies such as CNN and BuzzFeed on the backend, utilizing a similar crowdsourcing business model to reCAPTCHA.
What does it take to get you to work for free? It could be a required field in a contact form or a free tool that allows you to learn a new language. If you can create a perfectly balanced system, providing a valuable service on the front end and leveraging user actions to gain valuable data on the backend, Do It! I am personally infatuated with the crowdsourcing business model and hope to find a way to incorporate it into my business.