RE: The Ideas about Anti-abuse alert. Power Up #lub100
@alexmove, @papi.mati, @alejos7ven, @aymbionts
I echo a lot of what @stef1 has said.
I was aware that an automated tool was being tested within a couple of communities but I did not expect World Of Xpilar to be one of them - I thought that the communities had signed up to be part of the "experiment" and so the testing phase which @alejos7ven mentions should not have reached World Of Xpilar until it was production ready.
I also agree with @stef1's suggestion that investigation should be carried out elsewhere before a user is informed that their content doesn't pass "the test" - perhaps using the output of the tool you've created to create a list of posts elsewhere with a "plagiarism priority" that you can subsequently review. So that the tool can learn without a repeat of what's happened today.
You're also going to encounter a lot of users that have their own blogs, websites, portfolios elsewhere that we've already validated (everybody who has a label within World Of Xpilar has been through a validation process before receiving their labels) and will subsequently have their content labelled as plagiarised when it's not. Something I believe you need to consider with any kind of automation.
I'm pleased to hear that you're listening to feedback and I'd recommend sharing how you plan to approach the automation side of things in particular so that you can gather some feedback before going live - Perhaps within the Steem Dev community. You'll then be able to point people towards a discussion / debate which will explain the approach you've decided upon.
You're travelling on a difficult road and will inevitably stumble along the way.
The testing version will be limited back to a few communities that agreed for that.
Whitelist, like I mentioned before, will be created and added to the tool in the nearest future. When we will be introducing the tool to the WoX community, we will contact with the admins to make sure non-steemexclusive but verified users are included on it.
Plagiarism up to 25% and non-plagiarized content won't be marked in any way so the whole process will be a bit more invisible.
Those improvements will be included in the nearest future. Then we will be trying to figure out how to make it all even more invisible.
What might work well when rolling it out to more communities would be to share the list via discord or in such a way that the community isn't notified initially. I'd certainly be interested in seeing a list of posts that are being highlighted - the top 5 plagiarism percentages for example.
I think that any notifications within a community would be good to mention community admins and moderators (maybe not all of them but perhaps a few) and then we can warn or mute users too.
Mentioning the admins/mods is a good solution. Decentralized responsibility of verifying, downvoting and... Only admins can mute the abuser
Making the list of the potential abusers on discord instead of automatic comments is perfect but it sounds complicated (I'm green though, maybe it's very easy and I don't even know that) - is it possible to make @alexmove ?
I think @starlord28 might know how to do it 🤔
I was thinking that could be more of a manual process initially which if it goes well, then becomes the live solution. So you use it as a way to build up trust in the algorithm and test it in a live environment.
Okay. I'm lost over here. You want to know if mentioning the admins and mods in the comment is possible or do you want to automatically tag the admins and mods in their discord? Or are you talking about the decentralized way of downvoting? If its decentralized way of downvoting, it would take a month or two for me to have a solution.
It was the ability to send a message to Discord - I've got something in the back of my mind from a while ago that seems to think that you know how to do this 🤔
In the future, we may well make automatic posting to Discord.
Now I would not want to deal with the topic of automatic publication in the discord, since this is a new functionality and it will take time and effort.
Instead of automatic comments under posts, I will offer another option:
The first is to manually send lists to discord :)
Or more interesting and automatic - to publish the same automatic comments, but not under the post being analyzed, but under mine or your post. In other words, I create a post on my "Post Analysis" blog and, if I find a suspicion, I publish a comment under MY post that I created. And there in this comment I write as usual: "Attention! We need verification! The post is such and such, there is a suspicion of such and such,% uniqueness of such and such, links to similar texts."
What do you think? @papi.mati
I appreciate you stopping by and for your comment. As always it is I have another settings of my long night shifts and had to go yesterday, therefore was no time for even reading. I am glad that you are involved and next time will get in touch with you first maybe that will sort out the stuff quicker. Cheers,
I'm not part of the active team but I'm watching and helping where I can 🙂
I'd be interested in your thoughts on this idea when you have some time - https://steemit.com/hive-192037/@coding-gorilla/0000000000000111-proposal-to-change-focus
I'm going to keep working on the site redesign for now so there's no hurry or pressure to read it - I'll be interested in hearing your thoughts when you have some more time (and sleep).