Thank you for this provocation, I share your concern. As a reader, it's
disappointing to find material that looks like a press release, and
intimidating to flag or edit without doing research into the editing
history and editors involved. A quick, "back of the envelope" calculation
I did recently shows an alarming level of paid editing, with 1,017 "Paid"
status disclosures among en.wikipedia editors' user pages [1], which would
amount to 1.4% of active editors if these numbers were directly
comparable. This doesn't begin to account for any of the undisclosed paid
editing that must be happening.
As a technical contributor, I can offer two concrete initiatives which
might be helpful. Neither is a quick fix, but they offer spaces of
resistance that we can build upon.
* The JADE project [2] will create a structured namespace for patrolling,
and a talk namespace for coordinating work. You can think of it as an
enhancement to the patrolled edit flag, where patrollers can provide their
judgment in a format roughly equivalent to ORES predictions. We'll
eventually use these judgments to improve our training for the ORES AIs,
and our hope is that JADE will be integrated into tools like Huggle, to
make communication between patrollers more explicit. JADE is available for
experimentation on the Beta cluster [3], and we can move to the production
wikis after we get some feedback from experienced editors, maybe after the
upcoming Wikimania.
* We've also started work on an AI model to detect paid promotional
editing, based on the overly optimistic puffery that's commonly
deployed.[4] I'm excited about this approach, and once it's active we'll
be able to make good estimates of the scale of the problem, the number of
editors and sockpuppets failing to disclose their conflicts of interest,
and the financial resources pouring in. I imagine this would give us a
better idea of what next steps to take.
Cheers,
Adam
Wikimedia Scoring Platform Team
[[mw:User:Adamw]]
[1]
https://tools.wmflabs.org/templatecount/index.php?lang=en&namespace=10&…
[2]
https://mediawiki.org/wiki/JADE
[3] For example,
https://en.wikipedia.beta.wmflabs.org/wiki/JADE:Diff/376901
[4]
https://phabricator.wikimedia.org/T120170
On Fri, May 25, 2018 at 12:00 PM Anders Wennersten <mail(a)anderswennersten.se>
wrote:
My main worry, during my daily patrolling, is if we
manage to neutralize
the bad editing (vandalism, POV pushing) or if the destructive editing
is slowly successfully degenerating the great content we have created in
our projects.
In todays Sign-post it indicates an accelerating rate of decrease of
admins on enwp, and some likewise tendency on dewp. Is this a sign that
the "good" powers are losing out to the "bad" ones?
I also seen a very passive response to two massPOV editing . One, on 35
versions, is related to Hans Asperger, to state he was a nazi doctor
(false, even if he was somewhat passive in some cases). Here dewp
reacted quickly and after a while enwp, so these articles are OK, but in
most of the other 35 this false info lies unchanged. Also I react to the
effort from GazProm promoting their propaganda article /Football for
Friendship / in up to 80 version, and where almost noone has neutralized
it.
Are we slowly losing the battle against the "evil" forces? And if so,
is then our new strategy (being good in itself) and the plan to
implement it all too naive? For example I like very much the ambition
to help out on areas in the world where Wikipedia etc is not
established, but would it be more correct to put effort in regaining
control of the very many Wikipedia versions, that is definitely
degenerating and we are loosing what has been done on these. (as a test
look at "latest changes" on some of the versions with low editing, it is
depressing to see that there often are more vandal editing, not being
undone, then proper new material)
Would it be most appropriate if we all in a 2-3 years effort
concentrated on getting (back) control on our material in our projects,
before we start efforts in implementing the strategy we have agreed
upon. Perhaps a number of paid admins, vandal/pov fighters, about as
many as there are stewards today, would be necessary not to lose out.
Anders
//
_______________________________________________
Wikimedia-l mailing list, guidelines at:
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and
https://meta.wikimedia.org/wiki/Wikimedia-l
New messages to: Wikimedia-l(a)lists.wikimedia.org
Unsubscribe:
https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
<mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe>