• 1 Post
  • 27 Comments
Joined 2 months ago
cake
Cake day: September 21st, 2024

help-circle





  • “I’ve definitely learned that gender classifiers are an unreliable and flawed technology, especially when it comes to trans people’s gender expression,” Ada Ada Ada said. “I regularly see my algorithmic gender swing back and forth from week to week.

    Says the person changing themselves week to week to fit different classifications?









  • BMTea@lemmy.worldtoTechnology@lemmy.worldInside Meta’s Palestine Censorship
    link
    fedilink
    English
    arrow-up
    26
    arrow-down
    2
    ·
    edit-2
    1 month ago

    Anyone who uses Arabic-language social media has encountered this. They used to ban you for just making reference to “Al Aqsa” (Arabic name for Dome of the Rock) because their algorithm deemed it terror-related. They banned the word “shaheed” (martyr) too even though in Arabic it’s commonly used to refer to loved ones who died an untimely death, even in accidents. It’s also a name, which is hilarious because a member of their oversight board said in an interview that after they banned the word one of her coworkers named Shaheed had to explained that this was nonsense. Researchers did an experiment where they ran pages that used uncontroverdial Arabic keywords that would get censored, then do the same for Hebrew (including #death_to_arabs) which were left up and even gained traction.

    You can blame Meta to some degree, but the chief issue are US federal institutions that use notices and scare stories aimed at making risk-averse firms shut down anything deemed anti-American (which essentially means anti-Israel.) Just recently they’ve been sending FBI agents to knock on journalists’ doors if they publish the leaked Vance dossier and give them a “friendly reminder” that it may have been leaked by Iran. Even when the journalists mentioned it in their reports on the dossier.




    1. They were also fined 2,500 USD each.

    2. The case against them that most relates to what you’re talking about is in Michigan. They’re charged in accordance to a Michigan statute that bans deterring voters through “corrupt means or device”, referring specifically to disinformation that the two individuals specifically engaged in and their stated goals. That’s a world of difference from having a social media platform whose policies cultivate a userbase that seeks to get out the vote for a candidate and whose owner uses as a platform to advocate for that candidate. The case is actually going to the supreme court because the statute may be overly-broad.

    3. You haven’t provided any evidence or compelling argument that what they or Musk do falls outside of 1A protection. It seems to me that you’re implying that media institutions with a slant towards a political actor or party during an election is violating campaign laws? Please clarify.

    4. Invoking 20511 implies you believe pro-Trump disinfo on X posted by thousands of users constitutes “intimidation” of prospective voters. 30101 makes the “X support for Trump constitutes campaign finance fraud” argument look ridiculous:

    (B) The term “expenditure” does **not include-

    (i) any news story, commentary, or editorial distributed through the facilities of any broadcasting station, newspaper, magazine, or other periodical publication, unless such facilities are owned or controlled by any political party, political committee, or candidate;