Whose Image Is That? : Copyright Evolution in the Age of AI
Copyright law has had to evolve as technologies have advanced. The advent of AI-generated images (e.g. DALL-E 2) and other content (e.g. GPT-3) raises new questions about copyright protection that legal and technology practitioners will need to address. In this article, a little context is provided to help illuminate these new questions.
Let’s imagine it’s 1999 and you’ve just completed an oil painting. You’re excited to share an image of the work and a little story about the creative process on a new blogging site called Blogger. You start by taking a digital image of the painting using the first digital camera, a FUJIX DS-X. Then you craft a blog post in Microsoft Word. You talk about the brushes you used, how you configured the camera to take the picture, and how you enjoyed using Word to write the article. Finally, you upload your image and article to Blogger and attach the copyright symbol, ©.
You hope that people will take notice of your blog post and respond. While waiting, you fall asleep and have a strange dream…
Get Verlyn Fischer’s stories in your inbox
Join Medium for free to get updates from this writer.
Your post generates multiple responses. The first is from the paint brush manufacturer who posted a “Cease and Desist” comment. Since you created your oil painting using their brush, they claim copyrights to your painting. They say that posting a picture of it is copyright infringement. That seems farcical and you ignore it. But then, Fujifilm posts a similar comment claiming that any and all images generated by their camera belong to them. As if things couldn’t be weirder, Microsoft posts a “Cease and Desist” comment making the claim that any written work created with their application belongs to them. This all seems unfair since you could probably have generated the painting with other brushes, taken a picture of it with another digital camera, and wrote an article with a different word processor. Why can these tools lay any claim to your creative work?
Happily, you wake up and brush this foolish dream aside.
Okay, so let’s now fast forward to 2022 and imagine this: over the last few days, you’ve spent hours installing and configuring a software program on your computer that will generate images based on text prompts. Also, you’ve been refining the prompts to generate something you’ve had floating around in your mind. It took a lot of computer time to get the image you wanted — plus you had to learn the “language” of the AI model. In the end, you created something that is unique and interesting to you. You want to share it but want to reserve rights to do other things with it in the future. You post it to Instagram and attach that little symbol, ©.
Not long after the post, the AI software company, which made the image generation program you used, sends you a note saying that they own the rights to any and all images generated with their software and that your claim to copyrights in invalid. You wonder, am I dreaming?
While copyright law is not settled on the point of authorship of AI-generated images or other content, it raises a few interesting questions. Is there a point where tools used for creation become so easy that the person using the tool is no longer considered the creator of a work? Can the creator of the tool claim authorship of anything the tool makes even when the tool is directed by someone else? If the tool’s creator could claim authorship, how could they enforce it? Could they know whether any given image on the web was created with their tool or not?
As the courts take up these questions, we can take comfort in the fact that a well-documented history of technical progress is available from which to draw perspective and insights. It is still too early to say whether the very nature of copyright protection will change or whether it will merely evolve as it has since its inception in the early 1700’s.
