Design and Evaluation of a Digital Educational Escape Game for AI Literacy

This study presents the design and evaluation of "AI Chaos," a browser-based educational escape game built in Genially to teach AI literacy to students in grade 8 and above. The central aim is to determine whether a digital escape game can effectively build AI knowledge and sustain engagement, and the research uses a pre-post quasi-experimental design with 181 students. The game was adapted from a planned physical format due to organizational constraints, a shift that shapes how its findings should be read.

The evaluation found a significant gain in AI knowledge, with pretest and posttest medians rising from 5.0 to 8.0 and a large effect size (r = 0.745). Attitudes toward AI, however, remained largely unchanged over the short intervention. Participants rated the game positively overall (M = 4.24 on a 5-point scale), with particularly high scores for teamwork, perceived competence, and enjoyment. Narrative engagement was comparatively weaker, and text density emerged as a common criticism. Prior knowledge proved the strongest predictor of both learning outcomes and flow experience, while gaming affinity and age showed no significant effects.

The study also reports a structural imbalance in session design: debriefing averaged 18.3 minutes against 39.8 minutes of average playtime, exceeding the recommended one-third ratio. The authors treat this as a design consideration rather than a disqualifying flaw, given the game's educational purpose.

The research draws on established frameworks, including GameFlow for player enjoyment and prior work on escape rooms in education, and situates the game within ongoing questions about how AI attitudes form and change. Its contribution lies in providing empirical evidence that a digital escape format can produce measurable knowledge gains among secondary-level learners, while clarifying the limits of short interventions for shifting attitudes.