One of my key arguments for the ethics of open access is based on the idea that humans don't get their ideas from nothingness, and that they're in fact based on a wide range of sources that teach them everything from the language they're writing to the concepts they using to the background for the idea itself. I make that argument here. A very similar argument is now being made by technology companies who want to be able to use all the content that's out there to train their AIs. And I agree with that part of the argument. What the AI does is really no different from what we do when we read and watch all that stuff through our lifetime to earn an education. Where we come into conflict is when these same AI companies want to commercialize and copyright the output. And here I agree with the courts that have said AI-generated content should not be copyrightable. There may be something original in there, but it is such a small fraction of the overall output that slapping a copyright on that output is essentially theft from the commons from which this output was originally derived. But that means that people opposing copyright for AI-generated content can't be copyright maximalists.
Today: 4 Total: 1150 [Direct link] [Share]
