Comment on Automatically hiding spam works

  1. It would be nice if you gave us an idea of what can induce the system to mark something as spam so we can do our part to avoid it.

    Comment Actions
    1. I think the problem with this would be that spammers would also be able to avoid the filter, thus making it less effective. As well, spam detection systems (I have no idea what AO3 is using) are rarely custom, so they might not know how it spots spam.

      Comment Actions
      1. Fair enough, I guess we'll risk it...

        Comment Actions
    2. Green signpost saying Adopt-a-Highway: Butterflies & Rainbows

      A lot of spam screening software use AI learning to come up with their algorithms, so there's probably not any set rules to avoid anyway. I think it's likely that extreme parody crack that is intentionally written to be nonsensical might be more likely to get flagged than other works.

      Comment Actions