WHERE DOES THE ETHICS LIVE? MECHANIC-CONCEPT ALIGNMENT AND THE COST OF PARTICIPATION IN GAME-BASED AI LITERACY OUTREACH

Abstract

Game-based activities are increasingly used in public artificial intelligence (AI) literacy outreach. However, limited guidance exists on how game mechanics can meaningfully convey AI ethics concepts rather than simply making an event enjoyable. This study develops a five-level Mechanic-Concept Alignment framework that classifies an activity according to what its rules actually require players to do. Each activity is assessed against three tests: whether play requires handling AI content, whether the rules model the target concept, and whether play shows participants how they themselves performed. The framework was applied to seven activities at a one-day AI ethics outreach event in Kedah, Malaysia. Ninety-eight members of the public attended, of whom 62 completed an evaluation form, producing 124 open-text responses. What participants remembered was shaped more by participation cost, meaning the time, effort, equipment, visibility and adult permission an activity demanded, than by how well its rules carried AI content. Of the 24 respondents who named a specific activity, 18 named one of the least demanding ones. The activity whose rules best modelled an AI concept was named by 8.1 per cent of respondents, compared with 1.6 per cent for activities that only displayed AI content. Comparing the approved plan with the delivered programme explained why. The activity carrying the strongest AI content had been planned with a loudspeaker but was delivered with headphones, which made it invisible and inaudible to passers-by, limited it to one user at a time, and required parental permission for children. Raising the AI content of an activity therefore tends to raise its cost of entry. To help designers keep ethics content while lowering barriers to taking part, the study offers a one-minute activity model, seven redesign specifications and six practical rules.