authentication meets digitization

It’s something of an old story by now, but I heard an amazing piece this weekend on Tech Nation about a novel process that uses CAPTCHAs, the distorted series of characters you confirm when a website needs you to prove you are human, to decipher scanned words. Luis von Ahn, the Carnegie Mellon computer science professor that co-invented CAPTCHAs, started to “feel bad” about all the time people were wasting solving the meaningless puzzles (something like 500,000 hours a day worldwide), and decided to try to use the human authentication method to do some good. Dr. von Ahn has created a method that uses CAPTCHAs to allow humans to translate the 8% or so of scanned words that optical character recognition (OCR) technology can’t recognize. Even better, he wants wants to put this technology into widespread for the benefit of mass digitization projects, and I was glad to hear him mention the Internet Archive’s Open Library in addition to Google Book. Dr. von Ahn is also working on audio CAPTCHAs as a means of increasing screen reader accessibility.