Fictional Worlds, Real Connections: Developing Community Storytelling Social Chatbots through LLMs

Online Community
DOI: 10.48550/arxiv.2309.11478 Publication Date: 2023-01-01
ABSTRACT
We address the integration of storytelling and Large Language Models (LLMs) to develop engaging believable Social Chatbots (SCs) in community settings. Motivated by potential fictional characters enhance social interactions, we introduce Storytelling (SSCs) concept story engineering transform game into "live" entities within player communities. Our process includes three steps: (1) Character creation, defining SC's personality worldview, (2) Presenting Live Stories Community, allowing SC recount challenges seek suggestions, (3) Communication with members, enabling interaction between users. employed LLM GPT-3 drive our SSC prototypes, "David" "Catherine," evaluated their performance an online gaming community, "DE (Alias)," on Discord. mixed-method analysis, based questionnaires (N=15) interviews (N=8) reveals that significantly enhances engagement believability SCs
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES ()
CITATIONS ()
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....