Designing Player-AI Interaction: Engaging Players to Understand AI Through Play

Designing Player-AI Interaction: Engaging Players to Understand AI Through Play

Designing Player-AI Interaction: Engaging Players to Understand AI Through Play

Jennifer Villareale

Abstract

"With the recent advances in Artificial Intelligence (AI) technology, people are interacting with a grow- ing number of AI products in many aspects of everyday life. However, compared to traditional interactive systems, AI-infused products impose additional challenges to the current user experience (UX), such as low interpretability and unpredictability, making it increasingly difficult for humans to understand these sys-tems. This problem has precipitated a necessity in the Human-Computer Interaction (HCI) community to investigate approaches to counter this issue. Among the fast-growing body of literature, HCI researchers have proposed using games and playful experiences to help people better understand AI. Yet, limited re- search offers design guidance on how to approach the design of games toward this effort. This dissertation aims to help fill this gap by investigating how we can design the interaction between a human player and an AI agent, i.e., player-AI interaction (PAI), in AI-based games to engage players in developing their “mental model of AI” through play. This work presents four studies. The first study explores the design space for PAI in the scope of Neural Network (NN) games to understand how designers currently structure the interaction. The second examines how to design PAI through three case studies. The third examines how players approach the development of their mental model during gameplay to understand how this process happens over time—specifically, examining how players make sense of an AI player during a competitive interaction. The fourth examines how players develop their mental models of an AI player in a cooperative interaction and then compares how competitive and cooperative interactions impact mental model development and accuracy. As a result of this work, this dissertation yields design guidelines on how to approach interaction design in AI-based games, which interaction best engages mental model development, and the tradeoffs associated with these design decisions."

Reference

Villareale, J. (2023). Designing player-AI interaction: Engaging players to understand AI through play (Doctoral dissertation). ProQuest Dissertations Publishing. https://www.proquest.com/openview/b7cf572cdab3c63f059659fe7d2d67ac/1?pq-origsite=gscholar&cbl=18750&diss=y

Keywords

Player-AI Interaction (PAI), Mental Model of AI, AI-based Games