Post #2224205

2024-01-16 12:49 UTC

@donty@mastodon.tetaneutral.net @martenson@mastodon.world That's the thing. I wouldn't even object to AI training on copyrighted material, same as I wouldn't object to a human studying that same material to improve their skills. But I would expect anything that comes out of it to be owned by the AI, not the company who made the AI nor the human who fed prompts into it. With the AI not being a legal person, everything it does becomes public domain. See https://en.wikipedia.org/wiki/Monkey_selfie_copyright_dispute for a similar argument being made for animals.

Replies (2)

@dfyx@social.helios42.de 2024-01-16 13:09

@donty@mastodon.tetaneutral.net @martenson@mastodon.world Of course this shouldn't apply to cases where AI just directly copies a copyrighted work. But that's nothing new either. If I hire a guy on fiver to write me a fantasy story and he hands me a copy of LotR, surely I can't just use that as if it were not copyrighted and blame him for all legal consequences. But those cases are rare and mainly happen when you explicitly ask it to do so. Most outputs are probably original enough to fall under fair use or similar clauses.

Open ##2224206
@martenson@mastodon.world 2024-01-16 16:22

@dfyx@social.helios42.de The "AI" here is just a fancy word for machine learning-backed programming. Monkey has way more personality. And personality goes a long way. ;)

Open ##2224207