We find that 1) acceptance of AI varies widely depending on use case context, 2) judgments differ between demographic groups, and 3) people use both cost-benefit AND rule-based reasoning to make their decisions where diverging strategies show higher disagreement.
