Using content to train AIs was probably not on the radar when the first CC licenses were developed. But today we have this new kind of use that isn't exactly reuse, isn't exactly copying, isn't exactly... anything. What we have in this article is Creative Commons dancing on a very fine line, one where they want to say that using content to train AI doesn't infringe on copyright, and at the same time want to say that this use "must be balanced with equally valid considerations to ensure sharing ultimately benefits the public." It really feels to me that they're punting this one. I get the feeling they really really want to endorse the use of open content to train AI, and are responding to questions are ethical considerations by saying essentially that such questions are out the scope of CC licenses.
Today: 1 Total: 120 [Direct link] [Share]
