Whereas, a downvote/flag scenario will perhaps cause a chill, Alice might just switch to a different time of day to do her curation, hoping Bob is asleep
Posts are open for voting for at least 24 hours so that won't work.
Voting bots are readily available such as the steemvoter service, and others can be created.
To me this strikes me as a solution in search of a problem, and one that potentially causes all sorts of conflicts and perception problems that may well exceed its utility.
If negative autovoting bots became widespread then it could be seen as an optimization. But it isn't even clear whether a simplistic "negative vote every vote by this user" would express the desires of voters anyway. What about "negative vote every vote by this user, except these sorts of posts, where I agree with the upvotes". Or negative voting only on downvotes (or only certain ones) but not upvotes. All of these and more can easily be expressed with scripts/bots, but not with a direct negative vote. And in all cases, this equally incentivizes buying more SP in order to increase influence.
It is just not clear to me that a blanket negative vote is a use case that would actually have significant use, and if not then it is more trouble then it is worth.
Another consideration is, would this vote negation even have any noticeable effect, or would it just add overall negativity to the platform!
For a simplified example, lets say that Bob and Alice are the only two voting, and as described they have roughly equal stake. Alice votes on chickens and Bob votes on what he deems worthy. The rewards pool will be split 50/50 between the two topics. Now if Bob sacrifices some of his power to block Alice... it's a wash. He cuts her power by 50%, cutting his own by 50% in the process. Now when they vote again the next day... the pool is still split 50/50.
I don't see how vote negation does anything more than the simple strategy of voting on what you like... and not voting on what you don't.