Only humans can own a copyright.  But when we have human-like technology creating art, does that change the analysis? What about the work that the AI tool trains on?

Artists draw on other artists' works to educate and inspire them to create new art. Generative AI relies on training data, which consists of prior artworks (often protected by copyright law). These prior artworks were possibly collected and used for AI training without the original artist's consent or even awareness. Does using art in this way violate copyright law before the AI tool even creates the new work?

Some claim this is a new "fair use" - an exception to copyright infringement. Others say artists should be able to opt out of their work being used for training data or they should be compensated for their work being used as training data. Others still think artists should be a joint owner of the AI generated work. All of these questions - and many more - are at the heart of the intersection between copyright law and artificial intelligence. 

As is often the case, technology is lightyears ahead of copyright law.  This article points to the invention of photography as a historical example.  First, photography was seen as not "artistic enough" to warrant copyright protection, but eventually was an accepted tool for artists, like a paint brush or pastel. The difference here is a camera cannot think for itself to compose a great photograph.  AI, on the other hand, is developing work on its own, by analyzing real-life artists' work. Time will tell how this all unfolds and how lawmakers resolve the various issues around copyright law and generative AI.