2026-10-19 04:30 UTC
RE: https://mastodon.social/@Zygos/117461869086312391
It strikes me that there’s a real Catch-22 here for AI vendors:
They want to argue that training an LLM on other people’s copyrighted work without permission is OK because the LLM and its output are transformative use.
But if it’s transformative, then the people doing the transforming must be responsible for this new, original work they created. No more “we’re just indexing other people’s sites!” If it’s misleading or defamatory, that’s their responsibility now.
Can’t have it both ways.
Replies (1)
-
@inthehands@hachyderm.io 2026-10-19 14:56
To be clear, re several replies: (1) The post above is not about •my• opinion on the matter. It’s about which legal arguments make sense. (2) It’s possible for it to be both: it’s a coherent argument to say that some things an LLM does are in fact literal enough reproduction to be copyright violation •and• to say that humans are responsible for the tool. (That is in fact my own position.) The point of the previous post is that while saying “both!” is logically tenable, saying “neither!” is not.