> Saying "we already have a definition" when it's not clear whether it's been considered whether that definition would interact with something which is new, is... I don't even know what word to use. Square? Stupid?
The word you're looking for is "correct". The definition doesn't change just because circumstances do. If you want a term to refer to "open source unless it's for AI use", then coin one, don't misuse an existing term to mean something it doesn't.
> If you want a term to refer to "open source unless it's for AI use", then coin one
We even have such term already. It's source-available. Nothing necessarily wrong or bad about it. It only requires people to be honest with themselves and don't call code open if it's not.
Part of the background for this entire dispute is that prior to the OSI's founding, "open source" was a generic phrase which was broadly understood to just mean "the source code is available". See many documented cases in https://dieter.plaetinck.be/posts/open-source-undefined-part...
So it's a bit ironic to argue that terms cannot be redefined, when that's already what happened with "open source" and what got us here in the first place. If OSI had chosen a novel term (e.g. "Sourceware" was one option they considered), they would have been able to trademark it and avoid this entire multi-decade-long argument.
The word you're looking for is "correct". The definition doesn't change just because circumstances do. If you want a term to refer to "open source unless it's for AI use", then coin one, don't misuse an existing term to mean something it doesn't.