Hmmmm. Optimal would be if it was made to be expensive when used in mass. If somebody sent a single lone take-down request for my work, in genuine mistake, I would not begrudge him at all, and would not seek to penalize them. To err is human.

But if somebody sent a take-down request for my work, and it happened because they had sent a million take-down requests to everything on the internet matching a string search, then I would begrudge them and seek to penalize them. They would have done so KNOWING that a huge percentage of them would be wrong, but deeming it acceptable since it's no cost to them.