AI 윤리 교육을 위한 대화형 게임 설계: 집단 의사결정을 통한 윤리 담화 지원 = Designing a dialogue-game for AI ethics education: Supporting ethical discourse through collective decision-making

다국어 초록 (Multilingual Abstract)

This study introduces The Desk: Dilemmas in AI Ethics, a digital game designed to teach AI ethics through gameplay. Grounded in the social intuitionist model (SIM) of moral judgment, the game encourages players to engage in intuitive individual reasoning and collective decision-making on AI ethics topics via newspaper desk narratives. The design incorporates four key mechanics to support moral judgment and reasoning: (a) dialogues involving three players, (b) distinct phases for individual and group decisions, (c) a variety of article tones, and (d) public feedback. Tested with six university student groups (N=18), gameplay discussions were analyzed using the Discourse Quality Index (DQI) and Epistemic Network Analysis (ENA). Findings showed that group decisions typically aligned with the initial majority rather than minority persuasion. While players often rushed to agreement once a consensus seemed reached, the presence of minority dissent encouraged more in-depth discussion. ENA revealed that groups with high DQI scores experienced more dynamic shifts in individual opinions and better integration between majority and minority views. In contrast, low-DQI groups showed isolated minority voices, little opinion change, and limited integration of ethical perspectives. The study highlights the promise of dilemma-based games for tackling AI ethics and provides design recommendations: (1) postpone early consensus to allow alternative perspectives, (2) promote argument integration for high-quality discourse, and (3) disrupt cognitive closure tendencies by introducing uncertainty.