CC0

The abuse in this platform is of such rampant nature and in such a large volume that its beginning to be apparent that the only realistic way of addressing it is through the automation of detection and asessment processes. In this post a few solutions to existing problems that have been troubling curators and cleaners for a long time will be presented, in future posts of this series new solutions to other problems will be presented as well as updates regarding the implementation and results of the solutions presented here.

The "I didn't know" Problem

One of the most prevalent problems cleaners face is that almost all the plagiarists caught present as a line of defense that they "didn't know" that they had to cite sources, so in many cases these plagiarists go on for months or years without being detected, posting one plagiarized article after another and being rewarded by curators that believe the content is original and when a cleaner finally detects the plagiarism the famous "I didnt know" line of defense comes up, "nobody ever told me I had to cite sources", "I was just sharing", and a thousand similar excuses. This often ends up with the plagiarist being given the benefit of the doubt, second and third chances, and when caught again plagiarizing only then blacklisted as well as other measures. Abusers then go ahead an get a new account and start the cycle again, plagiarizing for months until detected and when detected again, present the same "I didnt know" excuse.

Now You Know

So the problem here is that the notification of the rules governing plagiarism in the platform are only given to the users when caught plagiarizing, the solution to this problem is to automate the notification of these rules and do so in an pre-emptive fashion, not only after plagiarism is detected.

The solution we will give to this problem will be implemented in cooperation with @themarkymark who has made a bot called @sources (See presentation of @Sources here). @sources presents a user with the rules regarding the citation of sources in various contexts when invoked with !sources in a post

The approach to follow is to automate the notification of every active user in the platform, adding language sensitivity to the bot so that for example when commenting in a post written in Spanish it replies in Spanish, and also adding an acknowledgement function to it so that the user can acknowledge being notified of the rules by replying to @sources a certain string like "yes I understand" so that the notification is not repeated. When the user acknowledges reception of the message it is then logged in a list of notified users that cleaners will be able to access, in that list the username will appear in one field, next to a link to the notification acknowlegement comment by the user and the date of notification.

In this manner when a user caught plagiarizing claims not to know the rules the username can then be queried by the cleaner and if the user was notified the claim of ignorance of the rules can be refuted and problem solved.

Background Checks

A while back this comment was made in a discord server by @bex-dk and it made us think about ways to facilitate the checking of backgrounds of users by curators and cleaners:

Recently in conversations with @r2cornell and @mariusfebruary -just to name a few- the same concerns were repeated, so we are presenting a way we found to facilitate this task.

There is a tool made by @inertia (presented here) thats seems ideally suited for this task, all you have to do is insert the names of two accounts and it shows you all the interactions between them, in this case we want to see for example the interactions between a certain user and @steemcleaners

and a certain user and @cheetah

or perhaps a certain user and @jaguar.force

You can even read the comments right there placing the mouse over the texts in blue.

One thing that could be done to further facilitate these background checks is to make one page where you enter the username to check and it opens several tags of the Steem-Togetherpage in your browser showing the interactions between the user and all abuse related accounts. We will develop such page with the help of @guiltyparties and present it to you asap, in the mean time you can use the method above.

Another way to check the background of users is to check if they are in any blacklist, a great way to do this is to use @themarkymark 's Blacklist Notifier which is a plugin for Chrome that readily allows you to see if a user is in a certain set of supported blacklists by just going to their profile. This is an example of what you would see in your browser using the plugin:

Feedback

What other tools do you think would facilitate your work as a curator or cleaner? Please let us know in the comments.

This is Jaguar Force,
Reporting directly from the Jungle.
Onwards!