Telling humans and computers apart through human computation
Original Publication Date: 2009-Oct-12
Included in the Prior Art Database: 2009-Oct-12
This articles describes a protocol for telling humans and computers apart through human computation.
(Completely Automated Public Turing Test to tell Computers and Humans Apart) is a type of a challenge response test used in computing to ensure that the response is not generated by a computer." (Wikipedia).
The common Captcha protocol runs as follows:
A computer server generates a problem
2. The problem is solved by the tested entity
3. The solution is passed to the server which grades it
4. The server concludes if the entity is a human or computer according to grade Major drawbacks:
1. Captcha tests are becoming increasingly harder (to counter the increasing success of spammers
and other interested parties in foiling them). Thus it is harder for humans to solve them.
2. Captcha tests are limited to problems that are easy to grade (for computers).
3. Captcha tests are vulnerable to man-in-the-middle / relay attack. References:
Captcha breaking: http://securitylabs.
Suggested here is the following protocol:
1. Entities request permission
It is assumed that the server has concluded the nature of some of the entities at this stage.
2. Entities are mated into pairs.
3. Each pair plays a game in which both entities prepare Captcha for each other and grade the solutions of the opponent pair member.
4. Several such rounds are played.
5. The server grades an entity by both the grade it gives to opposing entities and the grades it give to known entities.
6. The server concludes if...