Community Wishlist Survey 2022/Larger suggestions/Identifying Lobby Teams

This proposal is a larger suggestion that is out of scope for the Community Tech team. Participants are welcome to vote on it, but please note that regardless of popularity, there is no guarantee this proposal will be implemented. Supporting the idea helps communicate its urgency to the broader movement.

Random proposal ►◄ Larger suggestions The survey has concluded. Here are the results!

Identifying Lobby Teams

Problem: The problem is how to prevent corporate (not individual) sock-puppets - where teams of employed (or conflict of interest) editors collude to steer a narrative. While Wikipedia has a strong ability to moderate and manage those efforts by individuals to distort, rewrite, or promote a particular viewpoint, it is not at all as strong when dealing with large corporate interests, especially when pushing a narrative that, at face value, may appear to be robust and peer-reviewed, but is actually politically or financially motivated. As one of the most visited websites in the world, where people come to find out 'the truth', there is an increasing responsibility to prevent political/financially motivated agendas.
Proposed solution: This behaviour could be identified using modern graph analysis tools based on edit and comment behaviours. Current sock-puppet approaches use tools such as stylometric conformance analysis, but I believe that additional tools could be implemented that look at cross-user behaviours, especially where comments and edits are often shared, following patterns of behaviour that indicate focus on specialised content. This is particularly noticeable where there are a group of editors who are aggressively defensive in their given subject area.
For example, of a potential positive hit, it is not unusual, given any area of expertise, for editors to agree most of the time - however, when they agree all of the time, then their relationship becomes suspect.
Who would benefit: Everyone who matters benefits from increased vigilance.
More comments:
Phabricator tickets:
Proposer: 20040302 (talk) 20:37, 10 January 2022

Discussion

I don't know if I'm at liberty to discuss it here, but this already exists. I'm going to archive this proposal. Thanks for participating in the survey! MusikAnimal (WMF) (talk) 20:44, 10 January 2022 (UTC)[reply]
See mw:User:Ladsgroup/masz * Pppery * _{it has begun} 23:59, 10 January 2022 (UTC)[reply]
The example you showed only indicates sockpuppet identification through stylometric conformance. What I am suggesting is different, and involves co-ordinated efforts being made by separate editors who are working for the same lobbying organisation. 20040302 (talk) 11:19, 11 January 2022 (UTC)[reply]
@20040302 Thanks for clarifying, and for Pppery for mentioning Masz which I wasn't sure was public knowledge :) Anyway, I worry this project might be too big for Community Tech, but now knowing that Masz isn't what you're looking for, I will unarchive this proposal for the time being. Later when Community Tech reviews the proposals we may or may not deem this one too large, but if it is, we'll be sure to move it to our Larger suggestions category so it doesn't get lost here in the archives. Best, MusikAnimal (WMF) (talk) 16:44, 11 January 2022 (UTC)[reply]
Upon further investigation I think it's clear this is out of scope for our team, but it's otherwise a fine proposal so I'm going to move it to the Larger suggestions category. Regards, MusikAnimal (WMF) (talk) 23:27, 17 January 2022 (UTC)[reply]
Yeah, I don't think this is in the pile of things the team has time for, and it would need significant fleshing out. --Izno (talk) 00:19, 17 January 2022 (UTC)[reply]
A big task, perhaps better suited for its own persistent team. But I would say this is one problem we will eventually have to solve (or shut down). The # of editors who primarily edit because they are paid by syndicates or as editing consultants has been rising even as total editors slowly declines. In some smaller wikis and some topic areas these may already be dominant forces. –SJ talk 23:45, 23 January 2022 (UTC)[reply]
I'm no expert but I wouldn't have expected detection to be the main COI/PAID/SOCK issue, but enforcement. We need WMF Legal to start sending scary letters to the corporations and individuals blatantly in violation of our Terms of Service, as proud declarations on their websites of their "services" to clients make abundantly clear. — Bilorv (talk) 20:31, 28 January 2022 (UTC)[reply]

Voting

Oppose I think we should evaluate content on accuracy & relevance, not on motivation of an editor. Libcub (talk) 01:16, 31 January 2022 (UTC)[reply]
Oppose I already have significant concerns with the morality of @Ladsgroup:'s process - the community can't judge its accuracy for itself, we don't know its error margin, how much of the review (in the sense of "this is how confident we are in the closeness") is done by the CUs and how much is automated and just told to them, and knowledge of its existence is generally pretty low (I knew prior to it being posted here, but only because I was told about it by a CU). More relevantly, the comment above is absolutely right, we don't need better mapping tech, we either need Legal to become significantly more aggressive on dealing with them or a public relations campaign to see if we can impose pressure that way...and I think the latter would only help them. Nosebagbear (talk) 10:08, 1 February 2022 (UTC)[reply]
Support Oh, yes, this is desperately needed. Strongly support. XavierItzm (talk) 20:22, 1 February 2022 (UTC)[reply]
Oppose KingAntenor (talk) 07:17, 2 February 2022 (UTC)[reply]
Oppose Potential witch hunting tool. - Darwin ^Ahoy! 02:05, 5 February 2022 (UTC)[reply]
Oppose oppose votes because of confident in detecting collab socks (block real users) Thingofme (talk) 13:16, 5 February 2022 (UTC)[reply]
Support --Ciao • Bestoernesto • ✉ 19:12, 6 February 2022 (UTC)[reply]