pop & funWikipedia1980s–1990s

CAPTCHA began as a way to keep bots out, and it ended up training AI and digitizing books

the gist

CAPTCHAs are tests designed to be easy for humans and hard for computers, and their roots go back to hackers of the 1980s and 1990s who wrote in leetspeak, swapping letters for look-alike symbols so keyword filters on forums would miss them. Commercial uses began around 2000, when idrive.com and PayPal used distorted text to block fraudulent sign-ups, asking people to retype words that software struggled to read. Google bought reCAPTCHA in 2009 and used the human answers to help digitize New York Times archives and Google Books. Later versions moved away from visible puzzles: they analyse user behaviour and device attributes, such as canvas fingerprinting, and show image challenges only to users judged high-risk.

so what

Each solved puzzle once helped machines read scanned text, which means the bot-blocking tool doubled as free labour for digitization and machine learning. The shift to invisible checks also means that ordinary users are judged by hidden signals about their devices, which raises privacy and access questions.

From Wikipedia, the free encyclopedia · text CC BY-SA 4.0

Illustrative photo · Pau Giner, CC BY-SA 3.0, via Wikimedia Commons