The Spectrum of Authorship - Leading Cases in Copyright and AI-Generated Work

What is Human Authorship?

Copyright law generally recognizes human authors as the creators of original works. While works created by animals or machines may exist, they typically do not qualify for copyright protection unless there is significant human involvement or creativity involved in the process.

Someone took a striking picture of Oscar Wilde and a company reproduced it without the photographer's consent. The United States Supreme Court held that photographs were indeed eligible for copyright protection because photography involves creative choices in composition, lighting, and subject matter. A photograph isn't a plain machine-driven activity. The photographer exercises skill and judgment. Both in capturing and presenting an image, essentially telling a story.

Human Authorship: Copyright law generally recognizes human authors as the only creators of original works. While works created by animals or machines may exist, they typically do not qualify for copyright protection unless there is significant human involvement or creativity involved in the process. This might change if AI gains full sentience.

Some Level of Creativity

Under current law, to get a copyright for your work, there must exist some level of creativity. But exactly how much human creativity is required? How much does the human author need to mix their labor with the resulting art?

A monkey took a selfie and the court denied it having a copyright because the monkey was not a human and therefore could not claim human authorship. The photographer who owned the camera was also denied a copyright claim because he didn't click the shutter button to take the photo.

The underlying theme here is that of comparing a monkey to a human. The same can be said about comparing AI with humans. Currently, machines aren't human and therefore cannot own a copyright.

The Spectrum of Human Authorship

Copyright law recognizes that not all works are created with the same level of human input. There are varying degrees of authorship.

  1. Sole Authorship: A single individual has complete creative control and contributes the entirety of creative expression and originality to a work.

  2. Joint Authorship: Two or more individuals collaborate and contribute to the creation of a unified work. Each joint author shares equal ownership and rights in the copyrighted work, unless otherwise agreed upon.

  3. Collective Authorship: Works created by teams, companies, or other collective entities. The collective entity assumes authorship rights and responsibilities.

  4. Work for Hire: The individual who created the work assigns rights to the employer who hired them to create the work.

The Case for AI

Under the principle of collective authorship, there is a possible argument for AI-generated art. A shared authorship between the original human art creators of which the AI was trained, the engineers who created the software, and the human who entered the prompts to direct the AI.

What exactly does it mean to be human? Does it require sentience? Does it require being born of another human as all previous humans were born? Does it require having a soul? Lines are blurring between humans and machines and they have been for a long time.

New technology is often met with backlash until enough humans adopt it to turn the tide. When Photoshop was released there was broad backlash. Artists at the time claimed that it could be used to create fraud. But then it was broadly adopted by photographers everywhere. Photoshop was seen as a tool that artists used creatively to alter and enhance photographs, mixing their labor with the machine in a creative expression. Is this not so far from using AI tools to create art?

Fair Use

Can a human contribute 1%, say a prompt, and the machine produce the rest? Yes. Can a human then claim a copyright to it? Currently no.

The concept of Fair Use has been successfully applied to the images used to train AI models. AI image generators like Stable Diffusion have argued that human artists get inspiration from hundreds of thousands of pieces of art when training to fine-tune their own creative expression. This is considered fair use and is not plagiarism. The machine isn't stealing from humans. Rather it is humans who are getting inspiration and creating art using AI.

This may change under the weight of two important lawsuits. Getty Images has launched a lawsuit against Stable Diffusion, arguing that the AI is unfairly using their work.

Sales, Sales, Sales

  • Cosmopolitan magazine collaborated with OpenAI and digital artist Karen X. Cheng to create the first magazine cover created by artificial intelligence.

  • The Portrait of Edmond de Belamy is an AI-generated portrait painting created in 2018 by the Paris-based art collective Obvious. The painting sold for $432,500 including fees, over 40 times Christie's auction house initial estimate.

  • Artist Jason Allen used Midjourney to make the winning art piece called "Théâtre D'opéra Spatial" winning first place in the Colorado State Fair's fine art competition.

Current Lawsuits

Kelly McKernan, Sarah Anderson, and Karla Ortiz launched a class-action lawsuit in January of 2023 because their names appeared in prompts used in the creation of AI-generated art. They launched the suit against Midjourney, Stable Diffusion and Dream Up for infringing on their artistic identities.

The AI companies will likely argue that they are creating art from scratch as directed by human input from text prompts. Personally I don't believe this argument will hold up for very long because it is quite clear that the puck doesn't stop at merely entering text prompts.

It gets muddy because the machine is trained on copyrighted material. But aren't all artists allowed to get inspired by others?

Conclusion

"Good artists copy. Great artists steal." -- Steve Jobs

"Great poets imitate and improve, whereas small ones steal and spoil"-- W.H. Davenport Adams

Although the quotes appear contradictory they both confirm that art generation, no matter one's skill level, requires learning from others. No one creates art in a vacuum completely uninfluenced by others. This is a process as old as time.

In this post we touched on the concept of human authorship, what it means to be human, the concept of creativity, the spectrum of authorship in copyright law, fair use and leading-edge cases in copyright law as it pertains to AI.