Generation of Dilemma-Based Narratives: Method and Turing Test Evaluation

Abstract

This paper describes the analysis of the narrative generation aspect of the Generator of Adaptive Dilemma-based Interactive Narratives (GADIN). This system automatically generates narratives that are focused around dilemmas to create dramatic tension. The dilemmas are based on those clichés encountered in many of today's soap operas. A story planner creates sequences of actions that lead to dilemmas for storyworld characters. A narrative generated by GADIN is compared to that which took place in a television soap opera. The stories were subjected to a Turing-style test, with 127 participants, which showed that the readers could not determine which narrative had been written by a computer. © 2008 Springer Berlin Heidelberg.