Good points... I have always preferred the solution of "building up" things, rather than "tearing down."
The whole issue opens (for me) the potential discussion of whether "flagging" should simply be an autonomous process SOLELY to indicate spam, clickbait, plagiarism, duplicate content and so on... clearly "abuse."
Separately from that, maybe an examination of whether bots should have the same degree of voting influence as a human curator actually reading the content and making a decision. Or maybe a bot vote is still "a vote" but a bot vote is only worth 10% as much "reward" as a human vote... certainly might encourage more human curation, and narrow the gap between "votes" and "views."
What makes these things inherently bad? Who gets to decide what constitutes "and so on"?
I think this is a bad idea. But even if I thought it was a good idea, im skeptical that there would be any method of implementation that was not easily circumventable. And even if such a thing was possible, im not sure that committing to what would likely be an endless programming arms race with people like xeroc and recursive is a judicious use of steemit resources.