On Tue, Feb 17, 2026, at 6:47 PM, Vinnie Falco via Boost wrote:
On Tue, Feb 17, 2026 at 7:51 AM Matt Borland via Boost < boost@lists.boost.org> wrote:
https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Co...
tl;dr:
AI-generated output isn't copyrightable. Prompts are ideas, not expression. Using AI as a tool doesn't affect your copyright. Modifying AI output can create copyrightable elements, but only the human parts. No new legislation needed. The interesting part they buried: identical prompts produce different outputs. That alone settles it. You can't be the "author" of expression you don't control. The rest is 41 pages of agreeing with this while being polite about it.
That doesn't address the other (the interesting) side of the picture: is the output of an LLM subject to the source copyright? The line of reasoning presented cannot explain that (it would be like arguing that it's okay to fence stolen wares as long as you cannot easily figure out where it was stolen from, or if you could substitute the good at random with an item stolen elsewhere). The problem arises as soon copyrighted material is used during training. Yes, it does lead to interesting grey areas, after all humans do similar aggregation in practice. However, there's a very big difference in effort and implied merit involved. To be frank an optimistic (?!) statement like "no new legislation required" seems to fly in the face of public opinion and common sense. Disclaimer: I didn't spell out the entire discussion, so it is possible I missed more relevant obvservations on this side of the question.