So what are the claims? For the AI to generate something, it is trained by exposure to an extraordinary amount of pictures, photographs, text, and whatever already exists on a topic. Based on technologies like deep learning and natural language processing, artificial intelligence is putting out answers and content based on other people’s works on the internet. Here is the catch! Is this an act of copyright infringement, one might wonder? And if we end up concluding that it is, who is responsible for it?
OpenAI made a statement saying that they consider training AI software with copyrighted materials to fall under the “fair use” legal shield. Based on this fair use idea, we are allowed to copy certain copyrighted work without the permission of the owner if our end goal is to either comment or parody the material. Generative AI developers are claiming that the use of copyrighted material by their software is “transformative enough” so that the output does not directly copy any personal work.
So far, it seems like they won this claim, but governments are still looking into the problem, with so many artists saying their work is being stolen. Some would argue that there can be cases where an AI can generate transformed artwork that resembles the original too much. We will keep seeing this debate between copyright infringement and the fair use counter-argument since there are no clearly defined clauses that tell us when a practice is transformative enough. But wait, there is more. A new issue arises, and this time is about copyrighting the actual output of AI software.
Now, most countries have policies that give ownership rights for an image, text, or any other creative endeavor to a real person who created it. But what if the content is actually created by an AI? Is the user of the software still eligible to gain copyrights for the piece of content?
There is no rule available to solve this issue right now. So it is still up for debate in the next couple of years. Some people claim that we can see AI software as a tool. Following an analogy, we can consider the user as a photographer who uses a camera. The software is the camera, and the developers are the company that created the camera. Based on this analogy, the user should be eligible for copyrights on the image, right?
Well… yes, but it is not that simple. We might consider the user's effort and contribution when thinking about giving a real person ownership rights for a piece of art created using an AI robot. To this day, we have seen people winning art competitions with AI-generated images that they claimed were highly processed by them afterward.
We also saw cases like that of the writer Kristina Kashtanova. She illustrated a whole story using Midjurney’s software and claimed ownership rights to the final artwork. She argued that the images were created based on her original text input and also that she had to crop, edit, arrange, and overall process everything. And this granted her copyrights.
As you can see, the subject is complex. We will keep seeing discussions based on the copyrights of AI-generated works as it is becoming increasingly used by everyone nowadays.
.webp)
.webp)
