How about requiring AI companies to pay creators for training rights? Alternatively, models trained on the commons must be owned by the commons. Right now these AI companies are trying to have it both ways: it’s The People’s Data for training on comrade but ownership is privatized.
replyPractically speaking, who is going to enforce such a regime? Do you really want to give Chinese companies such a huge competitive advantage, that they aren't subject to the same costs as western companies? How do you even sort out which "creators" are owed, and how much? It's next to impossible, and would drown the legal system in litigation; it would likely cause more problems than it solves. On top of which you can find open weights for most, if not all, of the scraped material already. If you make those illegal to use, or prohibitively expensive, you just destroyed local LLM legality, and put the technology firmly in the hands of only the monopolists.
reply> What do people imagine can be done about it at this point? Offer a concrete suggestion.
replySimple. Free the companies from copyright liability, but after X amount of time they are required to release everything into the commons. The weights, the training scripts and the full training data (appropriately processed so that it can only be used for training and not for people to easily pirate whatever works were used). They'd still get a monopoly on their model for a little bit to recoup their training costs, but in the end would be forced to give back what they took.
I'm sympathetic, since I think copyright laws are far too extensive and generous. But it's not simple, there are a lot of companies that won't fall under your jurisdiction, and the question is if that will give them a competitive advantage that kills the industry for you, and ultimately costs you more than you gain.
reply