That's orthogonal. I don't want the app/website to know (and make decisions based on) whether or not it's being blocked or not... We can/should do IO/compute blocking in addition to faking input.

From an economic point of view, I understand the OP's point: it's a better deterrent to know know your data is bogus, than to be in the gray area where it may be right to use it or not.

If data is shit, approach dies, because it's not economically viable. If it's just unethical, it will never stop.