It will be pretty cool if there will be a coded bot that will scan effectively for abuse. However, this may consume a lot of effort and perhaps computing power? This may explain why no one is bothering to automate it.
Still, I hope people here can self-regulate the way the devs assumed them to be. Unfortunately, that's obviously not happening. So, uhmm, I guess this idea should really be tested.
Perhaps but it also require a lot of Steem power, loss of reward etc. Spreading the cost widely to many whales has many benefits and then also having @benjojo's idea incorporated will give it some sensitivity when required.
This is what i am hoping for as it is a low-risk way to try many various adjustments that perhaps can be later implemented into hardforks.
I also agree that people tend to not flag due to fear of getting back at, so I really think a 3rd party bot will be best. Fingers crossed for this to come into reality even at beta.