I posted some quick thoughts on generative AI on a closed Facebook group for the Alliance of Independent Authors. Then I tried to find them and, WOW, is searching a Facebook group a miserable experience. So here they are, in case I ever want to find them again:
I think the philosophical question here is interesting, but I have to admit I’m more concerned about the fact that concentrating any form of craft in AI concentrates power with those with the resources to scrape the web for vast amounts of labeled data, engineer the models themselves, and rent or purchase the hardware needed to train and run it. It feels like “democratizing art” or “democratizing audio” when we’re all producing AI-generated covers and audio for free or cheap instead of paying professionals… but it’s these Silicon Valley organizations who decide what we do or don’t pay. This is not a prediction of any kind, but note that over the long term there’s room for a Wal-Mart play: Run at a loss for a while, squeeze the smaller practitioners out of the market, raise prices when you’re the only game in town.
Maybe that doesn’t happen because the barriers to creating high-quality generative models become low enough to allow a robust competitive ecosystem… or for any number of other reasons. But we are already so reliant on a small number of very large tech companies for so much of our business (and our lives in general, outside of publishing). Right now, we’re lucky to have a huge community of cover artists and narrators who compete on quality and price. If that community is replaced by a few big tech companies, that competition may be gone. (This is also not a prediction; I’m writing this post on an Apple device on a social network run by Meta; obviously Big Tech has done some good in my life! But that doesn’t mean I trust them to prioritize me over their shareholders.)
“The philosophical question” is whether training an AI on Picasso’s paintings is different, IP-wise, from a person getting inspiration from Picasso’s paintings. It’s an interesting question! Seth Godin thinks there’s a simple answer. He might have the right answer, but if he does, I don’t think it’s because it’s simple.
I think it was in eighth grade science class that I learned about the square-cube law? Scale changes things. And the law (the common law, not the square-cube one) isn’t just a set of principles. It’s for protecting people, it cares about reality and changes in reality. If we decide machine learning is the same as flesh brain learning for IP purposes, we should do it because it’s the best decision for people, not because we’re pretending they are anything alike in how they work; they aren’t.
If you’re enjoying my writing, you can get some of my short fiction on your e-reader for the low, low cost of $0. Remembered Air is a collection of six poems and short stories not available anywhere else. Download it here.