In the spirit of attribution, credit as given by @downes for sharing this story came from @clintlalonde in Mastodon.
Naturally there is desire for some kind of concrete “rules” on use/copyright of content via and made by generative AI. On reading this article I’m left with a feeling that people want to have this court decision so they can weild copyright as a tool to protect the content they created, e.g. to control/commercialize it’s use- what is the creator of “Suryast” aiming to do with an assertion of copyright? Why do they seek copyright?
All of this seems almost inside our what I like to think we want to do in openness, not to draw fences around, but too make available. I’d guess the answer is, to apply something like a Creative Commons license, which is a copyright statement, one has to establish authorship to be able to apply a license.
Hence the fuzziness we have with non-human created content, right?
This also makes me think of a recent workshop I was in where participants made use of genAI for imagery. There seemed to be some interest from participants in wanting to assert that the things they made with genAI be excluded from AI training-- which struct me as odd since they had learned in the workshop that most genAI is able to do what it does by the rampant use of source material from others.
Did I mis-understand what the intent was to “exclude” their works from AI training? And all of this seems almost moot as how does one ever know what is in and what is not in a training set? It seems opaque.