Activity

I should say, the issue of LLMs isn’t entirely clear cut since they don’t actually redistribute any text. So their output may not be a copyright violation in the original sense. Could maybe be a derivative work of the training data though (see #1322).

There are a lot of open legal questions about AI. See https://hawleytroxell.com/insights/how-i-really-feel-about-chatgpt-from-an-ip-lawyers-perspective/. For example:

Copyright owners and patent holders have no recourse against infringing, illegal AI output since the law has not yet caught up to create a remedy. So if I ask ChatGPT to write me some Star Wars fan fiction and I then place that content on the internet or sell it on Amazon, Disney has no remedy—except to sue me somehow, because they are Disney and have a lot of money.

And:

I cannot register copyrights in content authored by an AI because I am not the author, and the AI cannot register its own copyrights because it lacks personhood.