Copyright in the world of victorious AI: the designer tries to protect the rights to the generated image

Allen did not deny that he used generative AI, and he was accused of accelerating the "death" of creative professions. But there was something more serious: he could not obtain copyright for his work.

The digital AI art enthusiast does not stop fighting. Recently, he filed a lawsuit against the US Copyright Office: the specialists of this organization did not allow Allen to obtain a patent for the illustration. What is happening now and how the situation may develop, we will discuss below.

Attempt to patent the image and refusal of the Copyright Office

After winning the competition, Allen applied to the US Copyright Office to patent his image. However, the office refused to register the rights, citing the lack of "human authorship" in the work. The commission then stated that it could not determine whether the prompts used were unique enough not to consider the work a generated image.

The Bureau's decision came as a surprise to Allen, as he was confident of success. According to him, he used 624 different prompts and spent about 80 hours on the work. The author made many edits to the final image, including adjusting the style, composition, colors, and tone. Allen claims that his contribution to the creation of the work was significant, despite the use of a neural network. It is hard to disagree with him: many of us remember what AI drew at the user's request two years ago. Without revisions, some of the pictures were so bad that they were hard to look at without tears.

As a result, Allen filed a lawsuit and went to federal court to challenge the registration denial. He insists that any work generated with the help of AI should be protected by copyright law, even if it was created based on just one prompt.

AI works, and all rights to humans?

Allen's story, as they say, "caught the wave" and became one of the reasons for discussions among representatives of the art world with patent offices and technical specialists.

In 2023, a group of artists filed a class action lawsuit against companies developing AI, such as Midjourney and Stability AI. The lawsuit claims that these companies used billions of copyrighted images to train their models. And all this without obtaining permission from the authors or paying compensation.

In August 2024, the court allowed the pre-trial process to begin, during which the parties must exchange evidence. So far, no final decision has been made on the main claims, such as copyright and trademark infringement.

AI companies claim that generated works are not subject to the law, as neural networks themselves do not generate exact copies of original works. In response to this position, the US Copyright Office is trying to understand the role of AI in modern realities. The staff has extensive experience: after all, the organization has been operating for more than 150 years. But it is not easy for them, as the Bureau has never encountered anything like this before.

In March 2023, the Bureau issued a ruling that works fully created using AI cannot be patented. But there are still controversial cases - for example, when the work is done using AI, while a person played a critically important role in the process. In such situations, the Bureau tries to find a balance between the right to authorship and the use of technology.

Now its representatives are studying the issue of the legality of using protected works to train AI models. It is important to determine whether this violates the law or falls under the doctrine of fair use.

Fair use is a legal doctrine in the United States. It allows limited use of copyrighted works without the need to obtain permission from the copyright holder. The doctrine was created to balance the interests of authors and the free dissemination of information. Several examples of its application:

1. Criticism or comments: use of the work in whole or in part for reviews or analytical articles.

2. News: use of copyrighted materials in news or reports.

3. Science and education: use of protected products for educational or scientific purposes.

4. Parody: creation of parodies of protected works.

Many companies developing AI claim that the results of neural networks fully comply with the provisions of the doctrine.

Théâtre D'opéra Spatial is copied by everyone

After Allen was denied copyright registration for his image Théâtre D'opéra Spatial, he faced a wave of copying of his work by anyone and everyone. Other artists began using his illustration without attribution and selling it on Amazon, Etsy, and OpenSea. Allen believes this is happening due to the negative reaction of the art community to what is happening.

Some of Allen's opponents have stated that they will deliberately pass off his work as their own or simply use it for any purpose because he will not be able to defend his rights in court.

Is there light at the end of the tunnel

There are certain shifts in the complex process of synchronizing the new reality with the legal framework. For example, in April, writer Eliza Shupe was able to obtain copyright for the structure and layout of the text created with the help of AI. Now the law prohibits the complete copying of her book, but individual fragments can be modified and published.

There are other cases as well. For example, the Bureau has registered more than 100 works that include elements created with the help of AI, such as song lyrics, written materials, and visual images.

Allen welcomes this court decision and emphasizes that uniqueness and authorship do not depend on the amount of time spent creating the work. Even if one prompt was used to generate the image, the final work remains an expression of a person's creative idea.

He now owns his own online gallery and sells physical reproductions of his works. Allen does not consider himself an artist in the traditional sense, preferring the term "digital creator." His goal is to create hundreds of illustrations with the help of AI, falling under copyright law.

Well, we have to wait for the court to conclude - then we will find out whether the servants of Themis agree with it or not. The decision will be extremely important for those who work with AI to create images, texts, videos, and other projects.