Can’t get any screenshots right now, but I left a comment on the recent Onion video about the white house trans alarm, just a funny little bit about how republicans can’t decide what bathroom trans people should be using.

It stayed up for a while, then just suddenly vanished. I tested a combination of words to see what triggered it. Any comments that I used the word “cis” in were instantaneously shadow banned. Additionally, when trying to let people know about that, I discovered merely referring to a shadow ban also gets your comment shadow banned, presumably to keep people from talking about the practice.

Just to be safe, I control-F’d and searched the entire comment section for the word cis. It was a video about trans people. It’s a straight up inevitability that it’d be used at least once. But nope, not a single reference to the word anywhere in the comments. That settles it in my mind.

I don’t know why I expected YouTube to be better than that.

  • Gaylactus@lemmy.world
    link
    fedilink
    arrow-up
    2
    ·
    3 days ago

    I’m always skeptical about these “YouTube is censoring X word in comments”. Can you please post the link to the video? And have you also tried to change the SORT_BY from TOP COMMENTS to NEWEST FIRST? YouTube will in a somewhat random manner, just “hide” some comments, for whatever dumb reason, from the TOP COMMENTS, but show all if you pick NEWEST FIRST.

    • Sombyr@lemmy.zipOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      3 days ago

      It was pointed out to me somewhere else that it probably isn’t the actual word being censored, but rather anything the AI they use to moderate deems as probably being intentionally controversial. The word “cis” does now show up, in one comment, and the word “cisgender” in another, both made several days after I pointed out the issue.
      For reference, and why I think the “AI removing controversial comments” is a plausible theory, the comment I replied to was “sorry, I had to use the bathroom,” which I replied to with

      “In the bathroom Matching your birth sex I hope 😤
      Unless your chromosomes are xx, then you better use the women’s room.
      Unless you also have a beard, then you better be using the men’s room.
      Unless you’re also a cis woman, then you better use the women’s room.
      Unless you’re also tall, then you better use the men’s room.
      Unless you’re also intersex, then you better not use either or we might have to acknowledge you exist.”

      And then after that all the comments of mine that got removed were pointing out that the reply had been shadow banned (editing to note: I did also try reposting the original comment worded differently and it still got shadow banned,) which also got shadow banned, until I switched to calling it “darkness forbidden,” where it went through. However, others were able to use the phrase “shadow banned” with no problem when replying to me.

      • Gaylactus@lemmy.world
        link
        fedilink
        arrow-up
        1
        ·
        edit-2
        3 days ago

        Right, this is why I’m always very skeptical about these claims, because I’ve seen people making claims, that are not general, and certainly not applied to everyone. I myself, have seen that happen. My top 3 hypothesis are:

        • There is an AI moderation happening, that most likely is targeting, not specific words, but rather specific manners of writing. I believe this is to prevent the huge amounts of bots that are constantly targeting the comment section, sometimes with scams, and have on several occasions become a very problematic issue. They still are, but it seems they come in waves… Like, imagine a new bot comes along, writes sentences on a specific pattern, and until that pattern is added to the AI moderation thing, they persist. Once they train the AI to pick up those new bots patterns of writing, the bots then poof, and in the process, may also target non-bots. For example, lately, one of the patterns of the new wave of bots is that they always end the post with 2 emojis. So it is feasible that the AI tool has learned that whoever uses 2 consecutive emojis, is a bot.
        • The YouTube comment section (and even YouTube as a whole), has been plagued with bugs, and it is a very old site, which is inheriting old code, and the web has dramatically changed in the last decade or so. So they are only doing patches over the code, and that is not always the best. There comes a point where it’s better if you just ditch the old code, and start something “newer”
        • YouTube relies on cloud services, all around the world, so there is a chance that the platform prioritizes syncing more important data, such has views, likes, and things that more directly apply to the monetization of the channel, which is a very tiny amount of data, when compared with all the comments. So it is plausible, that some comments are either being dropped to “save” bandwidth and data transfer, or synced at a later time, because they aren’t relevant for YouTube’s business model, which is, sell advertisement.