I’m not a lawyer, but my understanding of a license is that it gives me permission to use/distribute something that’s otherwise legally protected. For instance, software code is protected by copyright, and FOSS licenses give me the right to distribute it under some conditions.
However, LLMs are produced by a computer, and aren’t covered by copyright. So I was hoping someone who has better understanding of law to answer some questions for me:
-
Is there some legal framework that protects AI models, so that I’d need a license to distribute them? How about using them, since many licenses do restrict use as well.
-
If the answer to the above is no: By mentioning, following and normalizing LLM licenses, are we essentially helping establish the principle that we do need permission from companies to use their models, and that they have the right to restrict us?
LLMs aren’t produced by a computer. They are produced with the help of a computer as a tool. Like a novel is typed on a computer and the computer has been a tool in the process of creating that book. You’d probably agree the author has the copyright.
It’s the same with LLMs. The companies need lots of programmers to develop the software that does the training. Scientists who figure out what numbers to multiply to get coherent text. More people to scrape as much text as possible and curate the datasets. It’s really complicated.
Sure in the end a computer does the calculations. But that doesn’t make it the author. Neither does your digital camera take your photos or your hammer tile your roof.