- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
cross-posted from: https://feddit.online/c/[email protected]/p/2115447/all-ai-generated-code-should-be-open-source
cross-posted from: https://feddit.online/c/[email protected]/p/2115447/all-ai-generated-code-should-be-open-source
Except generated code in a lot of places simply cannot be copyrighted for starters (there is degree this can vary from place to place, from having some edits to being just a small part of the entire project)
That code is always covered by an explicit notice that tool output is not GPL, which cannot be used as a defence here
Source? I’m pretty sure that’s true for images, but copywrite law can vary greatly depending on medium, and context.
What they’re saying is true in case law, which is hardly settled or the same across US districts, but has some nuance.
The actual code that comes out of a language model is not eligible for copyright protection. However, if it is used as part of a larger work with significant human contribution, that larger work can be protected by copyright with the exception of the AI generated sections of the code.
So you could, technically, legally use the generated code but you would be in danger of having to explain to a jury of non-technical people how it’s possible to know that the sections you used are not protected when the company will contend that it was. It’s a big risk if the company is litigious and rich.
Threshold of originality doesn’t distinguish between images and code. Either it’s human creative work or it’s not. Chore isn’t creative work, so the amount of invested time doesn’t count.
Fair ask, here https://fsfe.org/news/2026/news-20260825-01.html