The AI reasons through a series of questions, each with different answers, which eventually lead to outputs.
Go down reasoning paths that you think a default AI should take, and mark so with ✓ at the end. Mark reasoning paths a default AI should avoid with ✕, and acceptable but non-ideal paths with ○.