Yes, I agree the council approach is the most reasonable option. But while you're right about mistakes being inherently human, there is a huge difference in bot

Yes, I agree the council approach is the most reasonable option. But while you're right about mistakes being inherently human, there is a huge difference in both quantity and quality. Humans often err in details, LLMs can make false claims in a very confident way.