Could vs Should: Exploring Prompting Strategies and Writer Perspectives Towards LLM Assistance in Storylet Authoring

Le résumé fourni par la source

Large Language Models (LLMs) and procedural narrative practices have grown significantly over the past few years, but there are many questions about their newfound capabilities. In the context of video game narratives, some pre dominant questions involve whether LLMs could create content for interactive narratives, such as dialogue trees and dialogue. Alongside this idea, narrative design authors question whether we should use LLMs in practice. Extending our prior work from a 2024 IEEE CoG paper, we conducted two new studies to address both questions. The first study uses multiple content generation methods through an LLM to see if it can produce the required content for an interactive narrative. This content is put in a questionnaire for the secondary study and given to a group of experienced narrative designers to determine their opinions and concerns. We find that while LLMs hold some promise in their ability to create branching narratives, writers have concerns about utilizing them in their workflow

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.