(Akatsuzi) (2023)
Image description: The image features a massive, ominous-looking robot that towers over two human figures and a car in a misty, forested landscape. The robot is predominantly red with visible wear and tear, giving it an aged appearance. Its eyes are large, round, and black with blue centers that resemble camera lenses. Wires hang loosely from its body, adding to the eerie atmosphere. The forest floor is wet and reflective, possibly indicating recent rain; yellow grass contrasts with dark trees. Mist permeates the scene, creating an enigmatic atmosphere; visibility is limited. The robot’s surface is scratched and worn; it appears not new, but rather something from another time or world.
There’s at least three Simon Stalenhag LoRAs that I can find. This person didn’t list all the resources they used because of a custom workflow they used, but it’s likely they used a LoRA at least.
I wonder if it’s feasible for an artist to have a LoRA trained and then sell that for people to create artwork inspired by them, maybe with a watermark, and then make it illegal to train a model on artwork you don’t have the rights to.
They can make problems for people who use their LoRA without license, but styles don’t belong to anyone. If someone is making art in a style like theirs, LoRA or not, there isn’t anything to be done.
I recommend reading this article by Kit Walsh, a senior staff attorney at the EFF if you haven’t already. The EFF is a digital rights group who most recently won a historic case: border guards now need a warrant to search your phone.
Yeah, I hadn’t really thought the idea through, I guess. I’m a recurring donor to the EFF, but I hadn’t seen the article, so thanks for that!
Edit: just finished reading it. It mirrors how I’ve tried explaining it to non-tech people in my life. Thanks again!