I saw some discussion on LinkedIn about applying the concept of fair use to AI systems like ChatGPT. Fair use is a legal doctrine under U.S. copyright law that allows limited use of copyrighted material without obtaining permission from the copyright holder. It applies to things like writing a review of a book or movie, or quoting a book. The assumption is that you’re adding value, and not using too much of the copyrighted work.
There’s clearly some sort of analogy between that and what AI does. It collects a bunch of content, evaluates it, and creates new content — adding value. Despite the similarities, I say the differences are more glaring. In fact, I think it’s more like a category mistake.
All the law about fair use assumes we’re dealing with humans. Applying a body of law designed for humans to a bunch of machines — which are nothing like humans in most respects — seems like a faulty application.
It would be like giving the right to vote to a TV character — or, better yet, to a character in a video game. It’s a misapplication of the concept.
There needs to be a new body of law around fair use that’s specifically designed for AI systems.
