An AI-literacy ethics scenario where two courses have opposite AI policies โ three decisions, real consequences, and nothing graded until the debrief.
Built to the eight evidence-based rules in the guide below, and the engine enforces them rather than just aspiring to them. Every decision probes the same competency โ whether you make your AI use visible before someone asks โ so itโs one moment of truth examined three times, not a string of unrelated choices. Nothing is marked right or wrong while you play; you get consequences instead, and the theory waits for the debrief.
The scenario itself is about transparency, not compliance. One professor permits AI and asks for disclosure; another forbids it without defining what counts. Youโre stuck at 11pm with a group deadline. The tempting choices are the rationalisations capable students actually use, the worst ending is fully reachable, and in it nothing bad happens to you at all โ the cost is paid quietly by a teammate, a professor with an unclear syllabus, and a classmate who stops posting.
Worth stealing: transparency, trust, and belonging are tallied silently and only revealed at the end, because showing meters mid-story turns consequences back into a score. They come back as a temperature reading rather than a grade โ cold to warm, plus how much each individual decision added or cost, so you can see what you built each time instead of only where you finished. The context is front-loaded onto the opening page so every screen after it is short and the decisions arrive fast. Illustrations are framed placeholders describing the art that belongs there, so the whole thing is playable before anyone opens a design tool.