A true random generator using human gameplay

Abstract

True Randomness Generators (TRG) use the output of an entropy source to generate a sequence of symbols that is sufficiently close to a uniformly random sequence, and so can be securely used in applications that require unpredictability of each symbol. A TRG algorithm generally consists of (i) an entropy source and (ii) an extractor algorithm that uses a random seed to extract the randomness of the entropy source. We propose a TRG that uses the user input in a game played between the user and the computer both as the output of an entropy source, and the random seed required for the extractor. An important property of this TRG is that the (randomness) quality of its output can be flexibly adjusted. We describe the theoretical foundation of the approach and design and implement a game that instantiates the approach. We give the results of our experiments with users playing the game, and analysis of the resulting output strings. Our results support effectiveness of the approach in generating high quality randomness. We discuss our results and propose directions for future work. © Springer International Publishing 2013.