Web appOpen in Telegram
WWikimedia General chat

Wikimedia General chat

✅ High trust
@WikimediaGeneral · group · Tech · indexed since 2026-07-15
1 261members−3 in a week
47writing in 30 days
484messages in 30 days
60 509messages in the index
E
Also, I have no opinions on whether this is good or bad, and I'm not agitating for either side of the issue. I'm just using it as an example for a broader trend, that shows that most of us don't understand the current "newbie experience"
  1. Z
    I did a preliminary analysis of block logs on English Wikipedia (2026 so far) and Croatian Wikipedia (2010–2020, described as project capture per WMF). Block reasons were grouped by keyword; "established editors" means accounts with 100+ edits. The most telling row may be "Broad or unexplained labels" for established editors: 18% on English Wikipedia versus 52% on Croatian Wikipedia, where the most common single reason was "ignoring instructions." A high share of vague or unexplained blocks of experienced editors might be a useful early-warning signal for smaller wikis. The "Not typical" row is less useful on its own, since routine username and vandalism blocks inflate it. Has anyone analyzed block-log reasons this way, as a possible warning sign of governance problems, for example the Trust & Safety disinformation team or researchers?
Whole thread · 2 replies →
  1. Z
    Yes, Kubura is the textbook case. What strikes me is how it ended. As someone put it on Meta at the time, Kubura "was banned coz of tax evasion (sockpuppetry on Meta votings) and not for being a mobster (keeping an entire local project locked)." Without that slip, the global community might have had no grounds to act at all.
Whole thread · 1 reply →
P
PseudomonasAny better ideas then?
Not forcing the newbies to use it would already be a very good start. And at the very least, test it minimally before bringing it to Wikipedia. Our newbies are way too valuable to be used like disposable laboratory rats by WMF devs.
A
Hi, Is there an abuse filter on your local wiki for UTM from AI tools? such as "?utm_source=chatgpt.com"
  1. P
    how exactly would it be a problem for Wikipedia searching something throught ChatGPT? 🤔
    1. A
      ChatGPT searches aren't the problem itself; they're just an indicator that the content might be AI-generated, which is what warrants monitoring
      1. P
        On my experience, that referral only appears when one is researching something using ChatGPT, not on article gen. Links used in article generation are usually clean of referrals. Só not sure such a filter would be extremely useful, if at all.
  2. W
    There is a filter on enwiki which flags instances of this specific indicator
    👍1
    1. A
      Could you share a link or the filter ID for that? I'd like to check how it's configured
Whole thread · 6 replies →
W
However, I will note that "?utm_source=chatgpt.com" alone is not enough to conclusively identify text as AI-generated, since all it means is that whoever added the citation found the website using ChatGPT
  1. A
    Fair point, but it often happens when ChatGPT itself generates the text along with the citations, and the user just copies and pastes the entire output. That's why it remains a strong red flag.
Whole thread · 2 replies →
P
it will probably detect correct use of AI by manually placing referrences on articles, and ignore potencially unreviewed articles or parts of articles created with it
P
could detect copy pasting of chat messages from a conversation, though those are usually easily detected just by reading them, as they are not on encyclopedic format

An open public feed from the search index ChatCrawler — “Google for public Telegram”; refreshed as the venue is crawled. Times are UTC.

Public content only, official Telegram API. About · FAQ · What we do not do · Remove a page · Catalog · Search · How we count