The widespread adoption of generative AI in education has created an urgent need to foster students’ awareness of AI-related risks beyond purely informational approaches. This paper presents the design, development, and evaluation of a rule-based narrative serious game implemented with Twine to support AI risk awareness and critical reflection. The game models everyday AI-related decision-making and deliberately contrasts the perceived convenience of AI use with its potential consequences, such as hallucinations and cognitive offloading. Following a design science research methodology, the prototype was evaluated in an exploratory quasi-experimental pilot study with upper secondary students and compared to a dialog-based RAG chatbot. Results suggest that while both approaches support short-term factual learning, the rule-based game shows indications of fostering a more differentiated and critical attitude toward AI use, whereas the dialog-based system primarily increases perceived safety and confidence.     «

The widespread adoption of generative AI in education has created an urgent need to foster students’ awareness of AI-related risks beyond purely informational approaches. This paper presents the design, development, and evaluation of a rule-based narrative serious game implemented with Twine to support AI risk awareness and critical reflection. The game models everyday AI-related decision-making and deliberately contrasts the perceived convenience of AI use with its potential consequences, such...     »