ChatGPT’s political responses have changed, according to a new study—earlier versions leaned left-libertarian, while newer versions show a rightward shift.
  • SnortsGarlicPowder@lemmy.zipEnglish
    9·
    3 days ago

    Always does. Didn’t twitter (before musk) make a chat bot that became a racist, holocaust denying, consiracy theorist bot?

    I think a lot of the early ones did. If you are scaping comments you are going to get a lot of racist shit.

  • Lad@reddthat.comEnglish
    29·
    3 days ago

    All of big tech seems to be. Gargling on those Republican balls and eager to please God Trump.

    • Possibly linux@lemmy.zipEnglish
      5·
      3 days ago

      Please make it easy for us to scrape content and also ban our competitors in China. Also, don’t pursue the antitrust stuff

      • big tech, probably
  • _cryptagion [he/him]@lemmy.dbzer0.comEnglish
    72·
    3 days ago

    ChatGPT isn’t shifting anywhere politically. It repeats what it hears, it doesn’t hold any opinions of its own because it isn’t sapient.

    People are shifting rightwards politically. Let’s not shift the blame onto an computer program, like it isn’t our fault the shit is like that.

  • Delta_V@lemmy.worldEnglish
    241·
    4 days ago

    I wonder if the trend we all noticed where it became progressively dumber over time was a result of an intentional lobotomy to make it more conservative?

      • SunshineJogger@feddit.orgEnglish
        2·
        2 days ago

        That it’s European.

        And it makes less errors with having data truly up to date. Also so far it feels like it hallucinates less.

        Other than that chatgpt is superior. Especially in contexual understanding and actually writing in a more human way.

        Lechat is somewhat more clinical in its answers.

        Overall chatgpt is better, though it’s American so I won’t use it anymore

  • irotsoma@lemmy.blahaj.zoneEnglish
    81·
    3 days ago

    That’s going to always happen when training data with the entire internet. The outliers will always skew thing more than the mainstream if the models are not designed to exclude them.

    And there are a lot of contributing factors, for example with right leaning stuff being more available for LLMs to process as the platforms are generally less concerned about privacy and more concerned about policing and control (that’s just what right wing is), of course the models are going to see more of it than the left leaning stuff where people are more on the repressed side, more likely to use more private communication methods, and less likely to be able to safely, publicly share outlier kinds of views to skew things the other way. Even the people who pretend to be extreme left-wing, like the USSR or the Chinese Communist Party, are usually, in reality, right-wing.