And it's precisely what democratic governance of AI looks like.

So, there’s a piece in Jacobin arguing that data center moratoria are a “terrible idea” making the rounds on social media and beyond. It’s pretty easy to see why this makes for some good discourse; naturally, there’s going to be frisson among AI optimists when a perceived opponent—here, the nation’s most influential socialist magazine—makes a case for aligning with the tech industry’s goals.

While I’m pretty unconvinced on all but one or two of the points that the piece itself raises, and I think it seriously misconstrues the class politics of data center fights, I do think it’s worth litigating this idea. Because I do believe we should be thinking about what a broader and more engaged politics of resisting, regulating, and ultimately governing AI might look like. It’s a good occasion, in other words, to ask:

  • Who is fighting data centers?
  • Why are they fighting them?
  • Are anti-data center movements a dead end—or a starting point?
  • Hirundinidae@beehaw.org
    11·
    2 days ago

    As others have pointed out, moratoria like the one Bernie Sanders and Alexandria Ocasio-Cortez have proposed are designed to buy time to work questions like that out—to develop good governance and regulation of AI—not install a permanent ban.

    It’s impossible to totally ban it since these models can be run locally. Anyway I think if we’re going to allow people to make AI slop and consume absurd amounts of electricity then there needs to be some guardrails put in place. Premium utility tax for electricity consumed (also apply this to crypto), watermarks for AI content, priority for renewables in new grid power, provisions for noise pollution. All of it.

    But like the article points out, what we need here is time and the tech overlords just don’t want to give us that.

    • BlameThePeacock@lemmy.caEnglish
      7·
      2 days ago

      Watermarks for AI content is a non-starter, there’s no possible way to enforce this. Pandora’s box has been opened, and even if you mandated all the big AI companies do it, the open source models that exist (or will be developed) will be available to bypass it.

      Nothing wrong with charging actual costs for electricity or expanding the grid with renewables. These are actually relevant ideas.

      Noise pollution? I haven’t heard that one before. As long as the data centers aren’t running generators, the noise from them should be lower than almost any other industrial user.

      • Jayjader@jlai.lu
        1·
        5 hours ago

        Forget charging “actual costs” for electricity, I think we should do like tax brackets and charge progressively more per kiloWatt-hour consumed. Let industry and corpo-sized services duke it out for who can consume the least, while protecting individual/personal use as well as small business, local-economy uses from the bigger actors throwing their weight around.

        Who knows, we might even be able to ensure free electricity to keep the lights on for everyone with the right kind of bracketing.

      • definitemaybe@lemmy.ca
        1·
        18 hours ago

        True that there’s no way to enforce watermarks, but it could be required to mandate that paid AI models include an invisible binary watermark in the last digit of the, say, red colour channel, so odd bits form something like a repeating QR code relative to the even bits; require the fans in keyframes in AI produced video, too.

        It wouldn’t work for text, obv, and this would be trivially easy to strip in post-processing, but it’s technically and legally possible, with low cost. Of course it wouldn’t affect local models, but not many people are running good image/video genAI locally anyway.

      • Hirundinidae@beehaw.org
        7·
        2 days ago

        They tend to have a massive amount of cooling units on top, which does create a lot of noise. There are also instances of the companies using portable diesel and even turbine generators to power their data centers, which has got to be infuriating to live near.