Anti-Piracy Group Takes Prominent AI Training Dataset ”Books3′ Offline

sapient [they/them]@infosec.pub to Piracy: ꜱᴀɪʟ ᴛʜᴇ ʜɪɢʜ ꜱᴇᴀꜱ@lemmy.dbzer0.com – 52 points –
torrentfreak.com

cross-posted from: https://kbin.social/m/tech/t/364852

Anti-piracy group Rights Alliance removed the prominent "Books3" database, that was used to train high profile AI models.

6

Once again it's that "worst person you know just made a great point" headline

I'm actually pretty pro-AI (and in particular, pro-FOSS AI), so I'm pretty unhappy about this myself ;p

If nothing else, this kind of shit will mean that only the existing "Intellectual Property" holders will have access to using AI. It would entrench things even more >.<

It may be illegal to distribute it, but it's not illegal to read it. Or train an AI on it, which is the same process.

Would commercializing the trained AI count as a commercial public performance though? The legal problems with AI don't come with the training, but when you start selling it.

I don't see what the problem would be. The AI model and its outputs are not derivative works from a legal perspective.