I’m Not a Robot: The Internet's Human Test

You’ve done this so many times you don’t think about it anymore. A box appears. You squint at some blurry letters, type them out, check the box. It takes about ten seconds.

You probably didn’t know that those ten seconds were going somewhere. For years, millions of people solving these security tests were quietly doing something else entirely. They were rescuing forgotten history that computers couldn’t read.

In 1950, Alan Turing proposed a test where machines tried to pass as human. Half a century later, a graduate student inverted it. The machine would do the judging. And the humans would get to work.

In this episode

• Turing's imitation game - the thought experiment that set the terms for AI
• Luis von Ahn and Manuel Blum - the Carnegie Mellon graduate student and his professor who built the wall between humans and bots
• reCAPTCHA - the internet security test that became the largest digitization project in history
• reCAPTCHA v3 - the invisible version

Episode Music

• James Opie / Nihilore, CC BY 4.0
  • "Whispers Invoke Paranoia"
  • "Do Not Look Back"
  • "Artifice"

Additional Reading

Pandey, K. (2022, July 25). History & evolution of CAPTCHA. Masai School. https://www.masaischool.com/blog/history-evolution-of-captcha/

Gugliotta, G. (2011, March 29). Deciphering Old Texts, One Woozy, Curvy Word at a Time. The New York Times. https://www.nytimes.com/2011/03/29/science/29recaptcha.html

Weintraub, S. (2009, September). Google acquires reCAPTCHA in two-for-one deal. Computerworld. https://www.computerworld.com/article/1331965/google-acquires-recaptcha-in-two-for-one-deal.html

Schwab, K. (2019, June 27). Google's new reCAPTCHA has a dark side. Fast Company. https://www.fastcompany.com/90369697/googles-new-recaptcha-has-a-dark-side

Transcript

Intro music

Speaker 0:20

You are sitting in the pale glow of your monitor late at night. You are trying to do something simple. Maybe buy plane tickets or access your bank account. You submit your password, but before you can proceed, you have to pass a test. A little box has appeared. Inside it, letters and numbers are twisted, warped, and smeared together like a note left out in the rain. Below it you read, "I'm not a robot." So you lean in. You squint. Guess that the strange loop is a Q. And the jagged line might be a five. You type it out and you check the box. Because you are indeed not a robot. But you failed. Another set of blurry letters appears. Back in 1950, the legendary mathematician Alan Turing proposed the imitation game. A test where a computer tries to fool a human interrogator into thinking that it is a real person. For decades, this was the benchmark of artificial intelligence. The human was the judge. The machine was the subject. But at the dawn of the new millennium, the machine started to judge us. I'm Daina Bouquin, and this is Lore in the Machine. The year is 2000. The internet is rapidly expanding, but it is being overrun by automated programs. Bots, masquerading as humans to spam message boards and hijack accounts. Luis Von Ahn, a 22-year-old graduate student at Carnegie Mellon University, and his professor Manuel Blum decide to build a wall to keep the bots out. Von Ahn had grown up in Guatemala City. When he was eight, he asked his mother for a Nintendo. She bought him an 8-bit computer called the Commodore 64 instead. He learned to program because there was nothing else to do with it. Von Ahn and Blum created a program called CAPTCHA. It stands for Completely Automated Public Turing Test to Tell Computers and Humans Apart. It is the imitation game in reverse. To convince the computer that you are human, you need to decipher distorted text that the bot's artificial eyes can't comprehend. The wall works, but every time you squint at those warped letters, it takes about 10 seconds of your life. And 10 seconds may feel like nothing, but with over 200 million CAPTCHAs solved across the globe every day, humanity was spending half a million hours a day typing gibberish. And Von Ahn finds this unsettling. He sees a vast, invisible factory. Millions of hours of our most precious resource. Human brain cycles. Being frittered away. He wants to harness this time for something meaningful. And it turns out the perfect opportunity was waiting inside one of the world's most famous newspapers. At the time, a massive effort was underway to digitize the historical archives of the New York Times. This was a staggering undertaking. The archive dated all the way back to 1851 and contained over 13 million articles. To do this, computers were using optical character recognition, or OCR, to scan and read the printed materials. But there was a problem. The old newspaper pages were worn, and the paper had degraded. The old ink was faded. The OCR software was failing. It simply could not read about 20% of the scanned words. So Von Ahn adjusted the CAPTCHA system to take those unreadable words from the Times archive and ask you to squint at them. Instead of one string of gibberish, reCAPTCHA started giving you two words. One was a control word that the system already knew. The other was a mystery, a scan snippet from an old article that had stumped the computer. Then the machine made a silent deal with you. If you typed the control word correctly, it assumed you were human. And if you were human, it assumed your guess on the mystery word was probably correct too. It took your answer, compared it to the answers of other humans passing through the same gate, and once enough people agreed, the mystery was solved. You didn't know it, but you were working, word by fuzzy word. You were helping decode the New York Times archive, rescuing stories from the 1850s that the machines couldn't read. Like an article about a pigeon shooting challenge in December 1865. Some local pigeon shooting champion from Jersey City wanted to take on England. These are details nobody thought to save but got saved anyway. One blurry word at a time. In 2009 Google saw the incredible power of this crowdsourced labor and bought reCAPTCHA. They plugged your ten second shifts into Google Books. An ambitious project to digitize tens of millions of books and create a massive library for the world. By 2018, over a billion people had unwittingly helped create one of humanity's greatest digital archives. We were tricked into building something wonderful. But we were also teaching the machines to read. So the test to see if you were human evolved. It wasn't just showing you twisted letters to make digitized books readable anymore. It started feeding you pictures of crosswalks, of traffic lights, and storefronts from Google Street View. By clicking those squares, you were training their image recognition systems. The machines needed to learn what a crosswalk looked like so a car could decide whether to stop. They needed to know where the landmarks were so the map could tell you where you are. The machines didn't just need to read, they needed to see. But then the images started to disappear too. Google released reCAPTCHA version 3 in 2018. And it doesn't ask you anything at all. There is no puzzle. No box to check. Instead, it watches. If you are on a site that has implemented version 3, it's invisible. But it sees you. It tracks how your mouse moves across the page. Whether you scroll in smooth arcs or jagged stops, the rhythm of your keystrokes, the way a human hand hesitates. It has seen enough of us. Billions of people performing their humanity over and over has taught it exactly what being human looks like. It no longer needs to ask. We proved we were human. And now the machines can too. I'm Daina Bouquin, and this is Lore in the Machine. If you enjoyed this episode, please rate and review and subscribe wherever you listen to podcasts. Thanks for listening.