Managing Bad Behavior Via Flags, Social Pressure and Stake
Recently we've seen a rash of spamming, trolling, and flagging from some large accounts.
This is all part of being on a blockchain with no rules and a lack of tools for Moderation except flags and criticism. Every society has a few uncontrollable bad actors who impact others' safety, enjoyment and in some cases earning. That is why most societies have a way of dealing with those people. Here in the USA, we have a jail system, one I think is broken and over-utilized, but it is a place to put people who can't or will not conform to the generally accepted rules of society.
The early vision of Steem was that stakeholders could manage bad actors using stake. As we have learned this sometimes gets difficult and really only addresses financial returns and visibility.
However, our large stake-holders seem to be reluctant to hold each other accountable and in some cases, the bad actors are so clever it is really hard to address their behavior at all. They have figured out how to make alt accounts that don't post or comment and are thus hard to fight in the current environment.
How do we preserve enjoyment and growth on the platform while also allowing for Free Speech and what I like to call Freedom of Stake? (it's my stake and I will do what I want with it) In the early days I thought we might get better at managing these issues and that flagging would become more common practice and more likely to be accepted. Rather the trollers seem to be getting better at how they attack those who they don't like.
In a current spamming event one user using several accounts managed to double the daily transactions, which might look good at first glance, but in doing that he brought several accounts to a standstill and numerous posts would not load or just completely crashed due to the antics.
So, what do we do? Some are even calling for blockchain level moderation from allowing one user to bloat the blockchain. At least we need better tools on the frontend.
It is a really challenging set of issues and any way that it is addressed, I suspect there will be some unhappy reactions...
What are your thoughts...
What should we do? If anything at all.
check out the posts on steempeak. It seems they've done some adjustments to hide all the spam comments. It will show hundreds of comments, but when you click load only the legitimate comments seem to show up.
Theres two completely different things going on right now. One we have this with the spam and need a way to control it. And I agree, we do. It's a major problem, and one thats going to remain and will continue to be a problem, probably a bigger one as more and more people get into the system.
And we are trying to expand it, hell maybe even bring some politicians in. But if they are barraged with so many comments in their posts that are just spam and flags, why should they stay? Why would they stay? We can't even get mad at them for not wanting to remain here.
A blockchain level moderation tool would go against the free speech and no censorship aspect that we are promoting won't it? It's a tough question, but one we must answer. Raise RC costs to a point where spamming isn't possible? But what about all the new people? What does that do to them? They can't interact either.
It's a tough thing, and we need to come to a conclusion and somehow stop the spam.
I think per account blacklists would be a good thing... Don't wanna see player x y or z? Anything? Block em! Problem solved. If anything looks off on any posts or analysis... It was the blacklist accounts likely down voting or otherwise. Give an option to allow people to inspect a players blacklist and easily add or remove to their own. Could all be done with json. ^>^ just getting apps to use this method.. that would be a hurdle, but a viable option.
I'm more than willing to discuss specific logistics on this method. Lots of things to think about.
But that sort of invites people to steal from the rewards pool from bots even more. Don't want some of the cleanup initiatives downvoting you? Blacklist them. I think more of a frontend solution would be good for now. If nobody can see the spam, whats the use of doing it?
Only for visibility on your frontend. Not on the block chain level.
Oh yea, it's a solid idea. Theres a mute button on steemit :). Mute is on a chain level, though I don't know much of what it does, assume its for frontends to use as they please, probably done the way you want.
I guess it's been awhile since I noticed. Hahaha. Totally forgot that was an option, never clicked it tho....
I've said from the moment that RC was introduced that it wasn't properly balanced. New users can't post more than once every so often with the delegation they get from Steemit, yet big users can spam the fuck out of everyone. The curve is too sharp. It needs to be made where it's harder for accounts to spam, and they need more and more RC to do so, but new users can at least interact to a normal degree.
RC is linear right? Why not have it be logarithmic like reputstion? But not as strong as that.
#sbi-skip
Posted using Partiko Android
It's a problem needing a mathematician and a lot of data to minimize the abuse while still allowing legitimate services. The minnows aren't nearly as big of an issue as the bots. But here we are so much time later and minnows still have RC issues.
I just delegated to a minnow the other day who was asking about RC, because I think they should at least have a bit of a better chance to start out. Even if they just turn out to be another one of the masses posting crap...at least they had a chance.
Definitely the case with helping newbies, I even working on a tool to simplify the process it for them (https://giftgiver.me, basically put in your username and as long as your RC are low enough and you aren't blacklisted, you can get 10 SP delegated to you). Without helping them, we can't grow and new users are what we need. But people aren't going to like whats going on with the spam and then realizing they can't interact themselves will frustrate them even more.
And thats true. Someone much smarter than me will need to look into the RC stuff. Hopefully we can help curb the abuse.
That's an awesome project. I had been thinking about projects myself to try to do delegations to minnows to help with SP, but I don't have the SP myself, so I have been trying to figure out how to deal with that problem as well. Haven't gone past the idea phase quite yet because it takes quite a bit. I guess it's likely that quite a few people have thought about this problem and what to do about it. Most are probably in a similar situation to me, where they aren't quite sure where to get the SP to delegate. Though perhaps with me it might have even more to do with not having that amazing idea that motivates me to start working on it. Best I can come up with is a tool to somehow let people auto-delegate to minnows that show worth somehow.
I shared info about your tool with a mod on steem.chat that often deals with minnows, suesa.
The solution is trying other front ends.
What?
Let me clarify. Trying other frontends won't fix flag wars. But it does help when dealing with comment spam. Steemit.com loads up comments by default. Steempeak.com does not as it gives the author the ability to auto-blacklist comments from certain accounts so that comment spam isn't automatically loaded for viewing.
Thank you for clearing that up!!
Posted using Steeve, an AI-powered Steem interface
I'm getting the idea, @whatsup, that there really is no cure. We either give up on freedom, decentralization, and the STEEM dream as we know it, or we just let things as they are and just allow anyone and everyone to be rolled over. Never mind who gets caught in the back and forth. It's the price we pay for true lack of centralized control.
Or, we could bring the fringes together and do some middle ground work. I would like to think that we could all agree that we are free to do what we want with our stake, as long as it doesn't steal or limit the use of someone else's stake (which means defining what all that means), and that crossing those lines has some kind of repercussions (whatever those are).
That said, whatever we do with our stake, while our choice to make, can also be subject to scrutiny, questioning, and in some cases, trying to stop that use by others, because of the way they exercise their stake. If we can somehow work in between those two parameters with little to no collateral damage, policing ourselves (as in, individuals do it to themselves) then maybe we've got something where freedom thrives, decentralization is still alive and yet we do spend some time thinking about what our actions are doing to others.
I'm open to suggestions. Really I am. I'm kind of tired of being told what can't be done by those who are supposed to be in the no, or who represent a certain philosophy or process of thought. This is not an impossible situation with no solution, and we don't have to be called dictators simply because we want a little decorum, common decency and civil discourse.
Maybe spam should be a problem for witnesses and developers to deal with. Witnesses could work with blockchain developers on rate-limiting parameters, and app developers could provide filtering. Then, the rest of us really shouldn't need to care about it. If users are seeing too much spam, maybe they should apply pressure to app developers to get the interfaces to deal with it better.
For voting abuse, I'm pretty sure I mentioned it to you before, but I still think something modeled after a 2nd price auction might be the best way to limit it. For example, Curation rewards could get paid out to all voters on a post, but the rshares from the highest voter could be sent back to the reward pool. This would mean that the highest voter would be paid according to the evaluation of all other voters, awards would only pay if two or more voters vote on a post, and a vote abuser would have to split their stake and throw away half of their voting power. It can't eliminate voting abuse, but I think it would curtail it, and with reduced need for flagging.
Google uses second price auctions for their adwords auctions, and they have the theoretical advantage that "bidding one's true value is a dominant strategy". Dunno if that dominant strategy would map into voting and rewards here or not, but it seems like it might.
It'll probably never happen here 'cause the current mechanism is already entrenched, but eventually someone might launch a similar platform that applies something along those lines.
If you implement too much control you will transform this place into another censorship place and freedom of speech.
Posted using Partiko Android
Agreed, it is a rough balance
"They have figured out how to make alt accounts that don't post or comment and are thus hard to fight in the current environment."
I got hit by that sort of random flag troll (for lack of a better term) the other day. The account had like 75 steem power and no vote power ( because it had been flagging all day) so it didn't hurt me too bad but it was a dick move none the less.
This wouldn't do much to address people with large stakes flagging for no good reason but I would totally delegate a little sp to an account that upvotes people who have been unjustly flagged by low power trolls to add more to the payout than the flag could have taken. The behavior would stop if the trolls saw that their flags ended up earning money for their intended victims. Let failure be their punishment, I say lol.
Perhaps we do need better tools. Perhaps a user should be able to control those that can reply to them in some way. Perhaps the owner of a post should be able to make certain comments less visible regardless of their flagging power. Maybe we should consider what certain users are doing with the Spam as a form of DDoSing or other type of intentionally malicious behavior that should possibly be addressed in some type of administrative way. Perhaps witnesses should vote to silence certain users from a certain front-end. Since certain users are causing Steemit to crash for other users, it can certainly be argued that their behavior should not be shown on the Steemit front-end. Perhaps witnesses should do like congress does and just do a sort of vote of disapproval that may be seen as a sort of warning.
The only solution I can think of after painfully watching this shit show unfold from the sidelines is a consensus-based solution around removal, and maybe some kind of penalty system.
In one case of the spam, I noticed the "exterminator" account has over -7 reputation. This is one of the accounts being controlled by the name we do not speak. That right there should be a warning sign that the account is probably trouble.
I think there needs to be some kind of system with manual controls in place who are alerted to accounts so negatively low in reputation (you must be doing some bad shit to get that low). Maybe it can be consensus-based to prevent massive abuse and targeted campaigns like the current flagging system. Reports are reviewed, maybe even Steem users have to "stake" STEEM to make a report if it's an accurate report, the STEEM is refunded, otherwise, it is kept as a penalty.
Ideally, I want to see the ability for witnesses to block and refuse to sign transactions on a node level. While it gets hard with the ability to create hundreds of bot accounts, maybe if we incentivised the community to report these accounts through a system like @steemcleaners then we can prevent their blocks passing through. Users could be rewarded with upvotes, delegation, STEEM or even a custom Steem Engine token.
More extreme cases like the present one, hard fork account removal. Strip the offender of their Steem Power and remove their account from the blockchain entirely. This is what some people in the community were threatening to do to Steemit Inc's stake by hard forking it from the chain.
Posted using Partiko Android
More filters!
Posted using Partiko Android