The Trump administration is reportedly weighing measures targeting Chinese AI models, from adding Chinese labs to sanctions lists to holding U.S. companies liable for security failures. Rather than impose an outright ban, Washington could use soft rules to deter adoption while protecting the market positions of OpenAI, Google, and Anthropic.
    • FaceDeer@fedia.io
      14·
      5 days ago

      Frankly, this kind of meddling just makes it more appealing to use Chinese models. Chinese models are often open-weight, meaning that they’re basically immune to foreign meddling - if China decides tomorrow to “cut you off” you just need to find some other provider that’s offering those models and buy access from them instead. Or if you’re a big enough organization, just rent or buy some hardware and run it yourself.

      This sort of meddling shows that the US government is eager to weaponize access to AI, and if you’re dependent on American models there are no other places you can run them.

      • melroy@kbin.melroy.org
        2·
        3 days ago

        Also the open models are not falling back to a worse model. Not like fable or sol are currently behaving.

      • eicker@lemmy.worldOPEnglish
        81·
        5 days ago

        This! The irony is that every attempt to turn AI into a geopolitical choke point becomes free marketing for open weight models. People tend to trust tools they can actually control. If access can vanish with a policy memo, resilience suddenly matters more than benchmark bragging.

  • CosmoNova@lemmy.worldEnglish
    8·
    4 days ago

    People seem to miss the point that it‘s probably more about information than anything else. They know this shit isn‘t productive. Deep down, they know they‘re selling crap. But it produces tons of data from users. Silicon valley knows this best. That‘s why they „protect AI companies“. They just want to harvest the data themselves.

    • eicker@lemmy.worldOPEnglish
      31·
      4 days ago

      There really is no better way to obtain data. And it’s not just data that’s fed into the AI, but also strategies, methods and processes: ideally, these are incorporated directly into the software.

  • _Nemo_@lemmy.mlEnglish
    7·
    5 days ago

    Good luck expanding that to the rest of the world which is just discovering that, while not perfect, China’s open weights models are

    a) good enough for 95% or all companies out there
    b) a lot cheaper, also because they’re
    c) self-hostable, meaning: they’ll always be available, unlike U.S. cloud AI which you’ll be cut off from whenever Diaper Donnie feels it’s time for yet another bribe.