When robots tell each other stories: The emergence of artificial fiction

Winfield, Alan F.

Abstract

This essay outlines a proposal for an embodied computational model of storytelling, using robots. If it could be built, the model would open the possibility for experimental demonstration and investigation of how simple narrative might emerge from interactions with the world, and then be shared, as stories, with others.

The model, as set out in this essay, has a surprising origin. It emerges from work toward making robots that can be safe in unknown or unpredictable environments (Winfield, 2014). That work takes the idea of robots with dynamic, continuously updating, internal models (of themselves and their environment) and links that with Dennett’s conceptual framework: the Tower of Generate and Test, leading to a new control system for safer cognitive robots. We then extend this schema, with the addition of a conceptually simple system for allowing robots to transmit and hence share parts of their internally modelled behaviour with each other. The core proposition of this chapter is that if we could build such a system, we would then have a model of robot-robot storytelling. That model might then be used to experimentally explore a range of interesting questions, for example narrative-based social learning or the relationship between the narrative self, and shared narrative.

• Publication Date: Jan 1, 2017
• Deposit Date: Jan 6, 2017
• Peer Reviewed: Peer Reviewed
• Book Title: Narrating Complexity
• ISBN: 9783319647128
• Keywords: artificial fiction, narrative, robots with internal models, robot-robot storytelling
• Public URL: https://uwe-repository.worktribe.com/output/904071
• Publisher URL: https://www.springer.com/gb/book/9783319647128
• Contract Date: Jan 6, 2017