4 ms·
It seems really unethical to train a model on someone else's words without their permission.
by krainboltgreene 3y ago
It seems really unethical to train a model on someone else's words without their permission.
- oasisaimlessly 3y agoNot to me.
- infoseek12 3y agoIt may be just about as unethical as someone discussing something they overheard in a crowd.
- krainboltgreene 3y agoSure if you have a really warped understanding of context.
- altdataseller 3y agoIsn't that basically what OpenAI did?
- krainboltgreene 3y agoIt's what the original dataset did, and then they further compounded on that.
- renewiltord 3y agoSounds horrifically unethical to try to censor people like you’re doing.
- krainboltgreene 3y agoThat's not what that word means and I'm not surprised you don't know that.
- valine 3y agoWhen I read your comment I trained my own mental model on your words. How is that any different? When a human reads words they apply a sophisticated theory of mind to contextualize the writing and the mental state of the author. If anything, LLM fine tuning is far less invasive than having a person read your writing.
- krainboltgreene 3y agoThis is an unserious argument and no one is swayed by it.
- valine 3y agoThe idea that reading a piece of text constitutes copyright infringement is ridiculous. Copyright isn’t some infectious thing. Reading copyrighted text doesn’t give the copyright holder a claim to the future creative work of the reader. You want to restrict model training, I get it. The debate is still ongoing, but I’m confident when these “copyright” claims work their way through court the AI companies will come out on top.
- krainboltgreene 3y ago> The idea that reading a piece of text constitutes copyright infringement is ridiculous. No man, it's not ridiculous. If I write a program that copies someone's book and try to sell it I'm infringing on that copyright. I cannot sell a zipped version of the Harry Potter books. I feel like there's so many people weighing in on this discussion who haven't actually done any real world copyright related stuff.
- valine 3y agoI see the source of your confusion. LLMs are not actually zips of the training dataset.
- krainboltgreene 3y ago