A moronic policy if true: 1. Copyright Office made it clear that AI generated output is generally not copyrightable irrespective of training data: https://www.

A moronic policy if true: 1. Copyright Office made it clear that AI generated output is generally not copyrightable irrespective of training data: https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence https://www.federalregister.gov/documents/2023/03/16/2023-05... 2. In fact, there is credible legal theory that goes as far as to conclude that training dataset license cannot effect the resulting weights under US law (EU's take on copyright make it less clear) 3. DMCA already provides Valve with legal cover in the unlikely event that training dataset license is somehow found to effect the IP rights of generated content. 4. By adopting this policy, they are acting more as a traditional publisher rather than a platform, thus exposing themselves to even more liabilities not less. This policy makes no sense to me no matter how I look at it. The harsh reality that Valve and everyone else needs to accept is that AI-generated content from "unclear" datasets is here to stay. People need to accept this fact^1 and move on. I already have. 1: Copyright is an incredibly limited, obsolete, broken invention that was never meant to handle concepts like this. It's very much like a poor analogy that we stubbornly insist on applying to situations were it simply does not fit. We will continue to find ourselves arriving to bizarre and nonsensical conclusions like this as long as we continue holding onto this broken invention. Reward authors in another way. Placing arbitrary limitations on information was never the right way to do it.