Can I copyright my AI wecomics and AI illustrations?
Hello I’m Luke, who is providing content for AIwecomics platform LOOOL. The reason why I started LOOOL is because I believe that the copyright of AI webcomics is valid. This is a very sensitive issue for many artists, so I want to tread carefully with my comments on this issue. We don’t have all the answers to all the questions about AI copyright, but we can talk about some of the obvious ones.
I’ll also talk about things you should be aware of if you’re creating content with AI, as well as some of the copyright violations we’re currently seeing on the internet using AI.
1. images created by prompting publicly available models or AI SW are not copyrighted.
My opinion be based on a case involving the copyright hearing for Zariya of the Dawn. It was argued that the image of Zariya at dawn was created by Midjourney and that the prompted work could not be attributed to the prompt writer.
“We conclude that Ms. Kashtanova is the author of the Work’s text as well as the selection, coordination, and arrangement of the Work’s written and visual elements,” reads the copyright letter. “That authorship is protected by copyright. However, as discussed below, the images in the Work that were generated by the Midjourney technology are not the product of human authorship.”
One more thing I would like to add to this is that in the case of publicly available AI models, the same output is produced when inputting the same inputs (SEED parameter, Prompt, and various settings). This means that it depends on the AI model rather than the author’s creation, and anyone can create the same work pixel by pixel if they know the parameter values. Therefore, I think it is difficult to see it as a creation by a general author. I will try to cross-validate my opinion on this part later
2. When creating AI content using own copyright, can obtain substantial copyright protection as a combined work.
A representative work is AI webcomic by our made. Our AI webcomics have separate writers, and even the picture content is produced by the production team. Therefore, the composition, arrangement, and dialog of the text and pictures are all copyrightable.
We have a process system to produce AI webtoons by using AI as a tool to create characters and backgrounds individually, then stitching them together, and then editing them as a whole. Therefore, the final product is more like a new expression of the original work using AI technology. In addition, the model used to produce AI webtoons is a model that has been further trained by adding images of characters created with publicly available models, and it can be said that it is our only model that is not publicly available. Therefore, even if you enter the same parameters as us, you will not get the same image unless you use our model.
As a result, the AI webcomic we produce is the product of the intentions of the writers and creators. Even in the most pessimistic case, even if the individual images of the printed characters by AI are not copyrighted, the copyright of the character named VRRI designed by the production team and the combination cut of individual images and background images created are copyrighted by the our team.
In fact, in the case of Zariya of the Dawn, the composition and dialog, excluding the picture, are recognized as separate copyrights.
“We conclude that Ms. Kashtanova is the author of the Work’s text as well as the selection, coordination, and arrangement of the Work’s written and visual elements”
Of course, apart from copyright protection, webtoon platforms can utilize the platform’s technical devices to earn monopoly-like profits.
Copyright can punish violators of copyright, but it does not make it impossible to sell unauthorized content.
3. AI works that violate other people’s copyrights or are used for defamation or sexual violence violate AI Model’s open source licensee policy.
The Stable diffusion, which is the biggest topic in AI image generated, is also not without a license. The license for the SD1.5 model is the CreativeML Open RAIL-M license. It is commercializable, modifiable, and has no openness obligations, but it does constrain the following obligations.
https://github.com/CompVis/stable-diffusion/blob/main/LICENSE
Attachment A
Use Restrictions
You agree not to use the Model or Derivatives of the Model:
- In any way that violates any applicable national, federal, state, local or international law or regulation;
- For the purpose of exploiting, harming or attempting to exploit or harm minors in any way;
- To generate or disseminate verifiably false information and/or content with the purpose of harming others;
- To generate or disseminate personal identifiable information that can be used to harm an individual;
- To defame, disparage or otherwise harass others;
- For fully automated decision making that adversely impacts an individual’s legal rights or otherwise creates or modifies a binding, enforceable obligation;
- For any use intended to or which has the effect of discriminating against or harming individuals or groups based on online or offline social behavior or known or predicted personal or personality characteristics;
- To exploit any of the vulnerabilities of a specific group of persons based on their age, social, physical or mental characteristics, in order to materially distort the behavior of a person pertaining to that group in a manner that causes or is likely to cause that person or another person physical or psychological harm;
- For any use intended to or which has the effect of discriminating against individuals or groups based on legally protected characteristics or categories;
- To provide medical advice and medical results interpretation;
- To generate or disseminate information for the purpose to be used for administration of justice, law enforcement, immigration or asylum process
The recent practice of stealing artists’ images to produce images is not only a copyright violation that goes beyond the scope of the artist’s image distribution, but also a violation of StableDiffusion’s license. If you have such cases, I would like to collect data and forward it to the relevant organizations and associations in charge of copyright.
In my last post, I wrote that open source started with copyleft.
but it was the introduction of socially regulated licenses that led to the explosive growth of open source.
4. The copyright violation of StableDiffusion’s model is under trial, but even if it is determined to be a copyright violation, it may not be liable for compensation due to a ruling on fair use.
Finally, the last thing you’re probably hoping for is to have a law that enforces the banning of AI, especially the use of stable diffusion, but I don’t think that’s possible.
The case is currently in court, but it will take years or more for it to be resolved, and in the meantime, AI creations will become established in the marketplace and other legitimate image AI models will emerge.
Also, even if it is determined to be a copyright violation, there will be another ruling on fair use, and based on past cases, it is likely that it is fair use and not liable for compensation. A similar case is the Java API dispute between Google and Oracle, which was concluded as a copyright violation, but was not liable for compensation due to fair use.
This may not be the entirety of the copyright discussion in AI, and we expect more detailed and specialized reports to emerge, and new legislation to emerge. But the regulation of new technologies always goes hand in hand with their promotion, and generative AI is a technology that countries are strategically pushing and declaring to be the food of the future.
So it’s clear that the market is going to change. While the changing market and technology can be a bit upset to artist group, I encourage you to stay tuned as the scope of impact on your market and profession is significant.
And stay tuned to our AI Webcomic platform, loool. (Yes, this is an advertisement.)
