Keeping the BOTS out. How figuring out who is human became lucrative.

 Jill Lepore’s *New Yorker* piece “Are You a Human?” is about the weird long history of proving our identity, from ancient tattoos to national IDs, to sci-fi clichés to corporate biometrics. But right in the heart of this historical trip is a fascinating look at CAPTCHA -- a tool that transformed an annoying security check into a huge, sneaky engine for crowdsourcing human effort.

CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) was created to combat the growing bot problem on the internet in the late 1990s by Luis von Ahn and a group of researchers at Carnegie Mellon. The goal was simple at first: stop automated programs from spamming forums, fixing elections, or carrying out cyberterrorism by making users decipher squiggly, distorted letters that computers couldn’t read.

The twist came when the researchers realized that all those millions of wasted seconds could be put to double duty. In 2007, von Ahn developed reCAPTCHA, which was acquired by Google two years later. And each time you had to decipher a weird snippet of text or spot a garbled word to get to a Web site, you weren't just proving that you were made of meat--you were secretly acting as a free digitizer for the Google Books Project. Your answers were fed directly into optical character recognition algorithms, transcribing scanned physical books that automated software could not parse.

When text decoding went out of fashion, the tests shifted to recognition of images. And clicking on crosswalks, traffic lights, and bicycles wasn’t just about security either, it was a giant, free training pipeline to teach machine-vision algorithms and autonomous vehicles how to see the physical world. In the process, internet users unknowingly offered up huge amounts of free labor to train the artificial intelligence systems that would eventually compete against them. As von Ahn himself would later reflect, it built a system that effectively vacuumed up human attention on a massive scale, ironically making machines so good that they could outsmart the tests themselves.

Classic CAPTCHAs are mostly obsolete now. At least in Lepore's view; I seem to run into them all the time and I hate them.  Supposedly bots have caught up, forcing companies to track mouse movements and behavioral data in the background, or turn to invasive biometric devices like Sam Altman’s iris-scanning Orb. But the age of CAPTCHA was the apogee of tech trickery: a security checkpoint that quietly transformed the whole web-browsing public into a legion of unwitting data labelers.