1307
New Study: At Least 15% of All Reddit Content is Corporate Trolls Trying to Manipulate Public Opinion (medium.com)
submitted 2 years ago by to c/technology@lemmy.world
215 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 21 points 2 years ago (3 children)

Yeah, looks like the default "word_wordnumbers" usernames that reddit gives you if don't change them.

I wonder if that would be an easy way to detect botting by not filling in that field for them.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 21 points 2 years ago

    Yeah, looks like the default “word_wordnumbers” usernames that reddit gives you if don’t change them.

    This change is when I knew Reddit was going down the shitter. Automatically handing out default usernames instead of requiring you to pick your own. The only people that could possibly help are a) people with absolutely no imagination whatsoever, b) bots, and c) people making a dozen alts to puff up their main.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 3 points 2 years ago* (1 child)

    Yeah, looks like the default “word_wordnumbers” usernames that reddit gives you if don’t change them.

    Funny enough, this started happening on YouTube (comments) as well, around the same time.

    The issue with sites starting to use numbering as part of the default username only started happening after AI posting became a thing, because an Achilles heel is the fact that AI can't come up with enough believable unique names for all the posts they want their AI bots to make.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 2 years ago (1 child)

    an Achilles heel is the fact that AI can’t come up with enough believable unique names for all the posts you want your AI bots to make

    That seems counterintuitive to me in the context of modern AI approaches. I'm wondering if you could elaborate on that a bit more.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 0 points 2 years ago* (last edited 2 years ago)

    That seems counterintuitive to me in the context of modern AI approaches.

    How so? Elaborate?

    I’m wondering if you could elaborate on that a bit more.

    This seems sufficiently explanatory to me, especially the italicized part...

    AI can’t come up with enough believable unique names for all the posts they want their AI bots to make

    Unbelievable usernames becomes an easy identifier/tag for identifying bot post.

    Edit: since this comment got downvoted (as the assumed reply) I thought I would elaborate a bit more.

    Basically, we name our user accounts to fit the society we live in's norms, it's naming conventions.

    If you just run a bunch of vowels and consonants together, that does not make a username, at least not one that people will recognize as a valid one created by a human being.

    Part of how bots are effective is in the quantity of bots that are used. Since it's near zero cost to spin up a new bot to make posts/comments, many can be made.

    However people can track the validity of a user name as being a bot versus human by the quantity of the posts/comments the username makes (only so many hours in a day, and human beings are busy with other things besides just posting on Lemmy), so no one single bot can make too many posts/comments at one time.

    Because of this, you need a large quantity of unique names, one for each of your bots, and they have to be believable ones by humans, so they're not identified as bots.

  • permalink
  • fedilink
  • source
  • parent