AI models can meaningfully sway voters on candidates and issues, including by using misinformation, and they are also evading detection in public surveys according to three new studies.

“It is not the case that misleading information is more persuasive,” Rand said. ”I think that what’s happening is that as you push the model to provide more and more facts, it starts with accurate facts, and then eventually it runs out of accurate facts. But you’re still pushing it to make more factual claims, so then it starts grasping at straws and making up stuff that’s not accurate.”

  • amemorablename
    16·
    1 day ago

    So this is the “manufacturing consent” setup for legislating against China-originating LLMs, right? Or have they already been doing that?

  • Commiejones
    15·
    1 day ago

    What are the LLMs gonna do, convince people to choose the wrong puppet of the bourgeoisie?

  • haui
    171·
    1 day ago

    AI will have zero additional effect imho. Elections in the west arent free nor are they fair. Forget it.

  • limer@lemmy.ml
    161·
    2 days ago

    I think it makes no significant difference, because elections are simply games and illusions, and the same paths will happen even if these effects from ai are a hundred times worse.

    At most, a scattering of candidates will change or be brought down, but same powers will remain entrenched

    • NuraShiny [any]@hexbear.netEnglish
      11·
      1 day ago

      I agree. People who are unable to recognize that the AI is giving them bullshit information will be fooled by it, but they would have been fooled by actual people telling them bullshit as well. It is highly stupid that people believe anything that AI says, but that is where we are.