This post originated from an RSS feed registered with Ruby Buzz by Christopher Williams. Original Post: The Reverse Captcha Feed Title: Late to the Party Feed URL: http://cwilliams.textdriven.com/xml/rss20/feed.xml Feed Description: Ruby. Rails. Stuff. Latest Ruby Buzz Posts Latest Ruby Buzz Posts by Christopher Williams Latest Posts From Late to the Party

Tim O’Reilly posed an interesting side-question on the O’Reilly Radar blog in a write-up on Google’s Image Labeler.

This is an interesting variation on the Turing test, in which humans generate and grade tests that most humans can pass, but current computer programs cannot pass. Is there another variation in the future, in which computers generate and grade tests that computers can pass, but humans cannot pass?

I’m not sure of the applicability of such a test, but the idea is intriguing none the less. I would think that there are plenty of problems that would be too difficult to solve by a human, but they are based on the assumption that a time limit is given for the production of a solution. You could simply give a very difficult mathematical question with a subsecond solve time.

One possible example without a need for a time-limit: a picture with a randomly generated steganographic message embedded and a text field with no explanation - computers fill it with the embedded message, humans would likely try to ascertain the image’s contents as the answer. Sort of a reverse captcha. The assumption here is that the algorithm for generating the random message is truly random, or not able to be reverse-engineered and the next value guessed by a human (sort of like those random card shufflers in Vegas that aren’t supposed to be beatable).

In any case, the idea is thought provoking. Anyone else have examples?