• RedstoneValley@sh.itjust.works
    link
    fedilink
    arrow-up
    2
    ·
    10 days ago

    What a weird article. They’re really trying hard to spin this.

    provincial government officials used ChatGPT to generate polarizing content and comments about the dire costs of data centers

    China has their own models now which are better at generating chinese propaganda than US models. This is just silly.

    The Chinese state is also influencing the training data for popular American AI models

    If OpenAI and Anthropic choose to crawl places where such influencial texts are found and then go on to include that into their training data it’s their fault.

    The researchers asked ChatGPT and DeepSeek, China’s top chatbot, the same political queries and found that DeepSeek’s answers were more favorable to China than ChatGPT’s 99 percent of the time

    I’m shocked, SHOCKED. well, not that shocked. You could rephrase that with political queries about the US and the answers would be biased towards the US as well.

    Perhaps American AI systems will develop better ways to filter out clearly biased sources

    LOL. The American billionaire class has no interest in filtering out bias, only opinions opposing their own. They’ve been openly embracing fascism for a while now.

    A technology that was meant to democratize information

    LOL. AI was never meant to democratize anything. They we’re stealing the world’s knowledge and now they are trying to sell it back to us, while making it harder to access to the original sources (this is the reason why you can’t find anything in google search results any more - they want to sell you Gemini)

    What a load of bullshit. I’m tired, boss.

    NINJA EDIT: I’m not saying China good, US bad. Neither can be trusted.