Learning AI Ethics through Game Design as Ethical Inquiry

This study examines whether facilitated participatory tabletop game making can help learners explore the AI responsibility gap, the difficulty of assigning clear responsibility when autonomous or learning systems contribute to outcomes that humans cannot fully predict or control. Fifteen students from computing-related disciplines at a UK university took part in a qualitative sequence combining researcher-prepared autonomous-driving cases, collaborative tabletop game design, prototype playtesting, and debriefing. The authors argue that responsibility judgements are context-dependent and revisable, and that making them into playable systems allows learners to articulate, test, and reconsider those judgements.

The findings show participants moving from broad statements about shared responsibility to more specific design questions about which actors could act, when, and from what position. Translating abstract views into rules exposed assumptions that had not been stated earlier, such as whether a role had the capacity to take a given action. Playtesting then returned these choices to discussion, most visibly when temporary numerical values prompted debate about whether they could legitimately represent differences in responsibility. Design difficulties, including disagreements and prototypes that failed to work as intended, also served as occasions for ethical reflection rather than only practical obstacles.

The study's scope is limited to one university, a small group of participants, and a single facilitator who also conducted the analysis, so the researcher's prompts are treated as part of the interpretive context. It reports observable articulation and reconsideration during the activity rather than measured learning gains, transfer, or longer-term outcomes, and it does not isolate the contribution of game design from facilitation or debriefing. Within these bounds, the paper concludes that game making offers a participatory and revisable form of ethical inquiry suited to problems that lack fixed answers.