Is anyone formulating prediction market questions asking AIs to brainstorm about edge cases in order to leave fewer of them uncovered by the market definition?

We do have humans brainstorming about such things, but this feels like something LLMs might be good at.