The AI Copyright Fight: A Guide

Whenever a client wants to skirt the edges of copyright law, the usual response of copyright lawyers is: "Do you really want to find out?" That is...do you actually want to learn how much you'd owe in damages from copyright infringement?

I heard this a lot from lawyers when I ran a niche indie publisher. I suspect it's a refrain that AI companies are going to have to get used to as well. Many AI image-generators, like Midjourney, Stability AI, and DeviantArt’s DreamUp, were trained on copyrighted images. 

In early 2023, the bill came due. Two parties—a group of artists (Andersen et al.) and Getty Images—separately sued AI image generators for copyright infringement. The lawsuits sparked discussion and debate among artists, engineers, VCs, AI companies, and the general public. I was surprised by how few people understood the nuances of copyright law. 

I thought now would be a good time to write an explainer piece for the tech audience on the legal protections afforded to creative works. I’m not a lawyer, so don’t take anything here as legal opinion or advice. My knowledge is based on my personal experience, research, and consultations with experts.

Copyright: A (very) brief history, from pre-Gutenberg to post-Google

Copyright got its start in an unexpected place: royal courts. About 500 years ago, European monarchs began doling out privileges and licenses to their favorite artists. No one else but chosen creatives were allowed to publish and distribute works of art. In other words, the role of the writer/artist and publisher were often the same. These publishers’ exclusive rights to print and make copies came to be known as “copyright.”