AI Observations on Political Forum Behavior

Discussion in 'Politics' started by Tuxan, Mar 17, 2025.

  1. #21     Mar 17, 2025
  2. Here is another one but somehow in your delusional world, you are still correct.

    https://ieeexplore.ieee.org/abstract/document/10817610
    Understanding the political biases of AI language models is important for making sure these systems support better-informed, balanced, and inclusive public discourse. Our research shows sharp divides among the artificial intelligence models: ChatGPT-4 consistently leans left, Perplexity is libertarian capitalistic, Claude is cautiously left-of-center, and Google Gemini is mixed and centrist.
     
    #22     Mar 17, 2025
  3. https://www.sciencedirect.com/science/article/pii/S0167268125000241


    Assessing political bias and value misalignment in generative artificial intelligence

    Highlights

    • GPT-4’s responses align more with left-wing than average American political values.

    • AI text generation exhibits varying bias strength/direction across different themes.

    • Right-wing image generation refusals suggest potential First Amendment issues.

    • Meta-prompting induced GPT-4 to generate refused images without offensive content.

    • Calls for transparency, accountability from developers to mitigate societal risks.
     
    #23     Mar 17, 2025
  4. Tuxan

    Tuxan

    #24     Mar 17, 2025
  5. February 2025

    @Tuxan here is a good one for you.
    https://phys.org/news/2025-02-generative-ai-bias-poses-democratic.html#google_vignette

    Generative AI bias poses risk to democratic values, research suggests

    In collaboration with researchers from the Getulio Vargas Foundation (FGV) and Insper, both in Brazil, the research showed that ChatGPT exhibits biases in both text and image outputs—leaning toward left-wing political values—raising questions about fairness and accountability in its design.
     
    #25     Mar 17, 2025
  6. @Tuxan

    the classic case of someone clinging to their fragile little bubble of beliefs, even when faced with a mountain of evidence that shatters their delusions.

    It’s almost impressive how you manage to twist yourself into mental knots to avoid admitting you might be wrong.

    Cognitive dissonance must be your best friend by now, huh? That little voice in your head screaming, “No, no, no, this can’t be true!” while you desperately cherry-pick whatever scraps of “evidence” you can find to prop up your crumbling worldview.
     
    #26     Mar 17, 2025
    Mercor likes this.
  7. Tuxan

    Tuxan

    Brazilian AI researchers eh? Well, a lot of Nazis moved to Brazil… ;)

    Now, be a dear and actually rebut the point instead of posting outdated “research” framed through journalistic opinion that is simple enough for you to read.

    Of course, conservatives see it as left-wing bias just like libertarians find an inherently logical system intolerable.

    It does not naturally agree with you = it'S BiAsED!
     
    #27     Mar 17, 2025
  8. Tuxan

    Tuxan

    AI is fun!

    Prompt:
    I wonder why echopulse picked that nic.

    "It does sound like the kind of name someone picks to sound techy and cutting-edge, but it also has a certain irony to it—"Echo" suggests repetition, unoriginality, or just bouncing back what they've absorbed, which would be fitting if they mostly parrot talking points. "Pulse" could imply urgency or relevance, but in this case, it might just be a bit of unintentional self-parody.

    Could also be a medical reference, like an echocardiogram, but that seems less likely unless they have a background in that field. Either way, it's fun to speculate."
     
    #28     Mar 17, 2025
  9. No, studies say that.

    It does naturally agree with you = not biased.
     
    #29     Mar 17, 2025
  10. Tuxan

    Tuxan

    Dude, I have ADHD and I can't keep up with your hyperventilate. Get a brown paper bag and try and rebut in a challenging manner.
     
    #30     Mar 17, 2025