Have you ever seen a computer-generated image? Maybe you’ve seen a photo and thought, “Wow, that looks so real!” But did you know that the computer might have made it?
Generative modelling is a fancy term that means a computer can learn patterns in a set of images and then create new ones based on those patterns. This is kind of like an artist looking at a bunch of paintings to get ideas for their own artwork.
One way computers do this is with something called GANs (short for Generative Adversarial Networks – don’t worry too much about the name!). With GANs, the computer has two “friends”: one that tries to make new images, and one that tries to tell whether the images are real or fake. The two friends work together to make the most realistic new images possible.
So, the next time you see a cool computer-generated picture, you can impress your friends by saying, “Hey, that might be the work of a generative model or a GAN!”
The question of who owns the copyright to a Generative Adversarial Network (GAN) generated image is a complex and evolving issue that hasn’t yet been fully settled by courts of law.
In general, copyright law usually grants ownership to the creator of an original work of authorship. However, with GANs, there are multiple programs and algorithms involved in creating an image, and it can be difficult to determine who the “actual” creator is.
Moreover, it is often the case that the original dataset from which the GAN creates images is owned by another person, complicating the question of ownership further.
At the moment, there is no clear legal precedent on whether GAN generated images can be copyrighted, and the issue is one that may require courts to grapple with as the technology becomes more prevalent.
However, do look up derivative art as that is different laws.
“Can I claim copyright for my pure AI image?”
“At present, no”
