The video gaming industry offers richer experiences through increasingly complex game mechanics, often hindering learnability. This research explores integrating an LLM-based chatbot, “Daisy,” in Stardew Valley, a narrative-rich role-playing game where learnability is critical. Over three weeks, 24 participants—14 new and 10 experienced players—engaged in a diary study and post-interviews. Analysis of diaries, chat logs, gameplay videos, and interviews revealed three themes: seeking information support, playing with chatbots, and addressing practical challenges. Findings show Daisy’s potential to enhance learnability through natural conversations, fostering immersion and emotional engagement, though issues like hallucinations and context awareness require improvement. This work highlights preliminary insights for integrating LLMs into narrative-rich games.
https://dl.acm.org/doi/10.1145/3706598.3713310
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2025.acm.org/)