@allpurposemat@mastodon.gamedev.place
Post #2732390
2026-05-18 07:38 UTC
Replies (3)
-
@muvlon@hachyderm.io 2026-05-18 07:58
@allpurposemat@mastodon.gamedev.place If you're talking about the Thaler v. Perlmutter case, the interpretation that they ruled AI output as "public domain" is a serious misunderstanding, widely reported by sensationalist media. All the courts did was decline to register copyright in the name of an AI "author". However: - Copyright registration is optional, copyright itself is enforceable without registration. - They didn't say anything about the status of the works in question, and definitely not that they're public domain. - It's my reading that this was thrown out on a technicality, because Thaler tried to have the AI be the legal copyright holder, which the law does not permit (it must be a human). It's very possible that the slop would have been copyright-registered had Thaler put himself as the author. In my view, the jury is still very much out on this, and I would caution against using LLM-created code as if it were public domain.
-
@noisytoot@berkeley.edu.pl 2026-05-18 12:24
@allpurposemat@mastodon.gamedev.place In the UK it is copyrightable and the copyright holder is whoever prompted it (assuming it doesn't plagiarise anything, of course, which it does)
-
@yoasif@mastodon.social 2026-05-18 13:17
@allpurposemat@mastodon.gamedev.place I wrote a post about this, but it is US-centric unfortunately: https://www.quippd.com/writing/2026/04/08/ai-code-is-hollowing-out-open-source-and-maintainers-are-looking-the-other-way.html I'll have to learn a bit more about how the Europeans think about this, especially if there is some special interaction with licenses that rely on US law.