AI Art? More like Copyright Nightmare!
Modern technology has hit a new milestone — artificial intelligence can now craft “art”. Simply open up a powerful image generator like OpenAI’s DALL-E, type in a prompt such as “majestic cat swimming in an underwater cavern,” and within 1 minute you have yourself a piece of art that you envisioned. Although this might initially sound intriguing, the veneer of innovation reveals a troubling reality.
Here is where I stand on this topic and matter: AI art can never be ethical as it goes against all concepts of artistic integrity, ethical coding, and model training. Not only are these AI models taking art from artists without consent, the fact of the matter is that the prompter and software creator should not be allowed to use AI art for their own monetary gains. Heck, maybe AI art should have never existed in the first place due to how difficult it has become to file any sort of complaint or formal lawsuit against its usage, especially when it’s employed discreetly or without proper authorization!
Considering the challenges in seeking legal recourse, the way AI artwork gathers its information is part of the issue. It utilizes a Generative Adversarial Network to collect extensive datasets that include visual content such as paintings, drawings, photographs, and more. However, this process of collecting datasets and training the model through complex matrices and equations results in a detached, algorithm-driven approach that lacks proper consent. Rather than fostering genuine creation, it appropriates the hard work of artists to generate a mix of repetitive, non-representative, and formulaic artwork (Elgammal).
This model signifies that the process of generating art has become more accessible and instantaneous, leading to a plethora of ethical considerations, particularly concerning copyright and intellectual property rights. With AI art, major tech companies such as Microsoft and Google no longer have to hire artists to bring their creative visions to life; instead, they can rely on these AI image generators to do all the heavy lifting for them.
Hence, one of the biggest issues with AI art is where it stands within current copyright laws.
As these models undergo neural network training, they gain the ability to analyze datasets containing artworks from renowned artists like Leonardo Da Vinci, enabling them to emulate specific art styles. In a hypothetical scenario where Leonardo Da Vinci were still alive, it is logically sound to argue that he would be entitled to full copyright of all artwork he created by his own hands.
However, this becomes tricky when talking about AI-generated artwork created using his style. According to the U.S. Copyright Office, once you create an original work and fix it — such as taking a photograph, writing a poem or blog, or recording a new song — you become both the author and the owner (U.S. Copyright Office). It also must meet the following requirements to gain protection under the Copyright Act:
- an original work of authorship;
- fixed in a tangible medium;
- that has a minimal amount of creativity
If the artwork does not meet all three of these requirements, it does not qualify for copyright protection (Mathur).
While this law appears to protect artists from people claiming that a certain AI art piece is theirs, this can become a double edge sword. Given the situation where someone were to re-distribute your work without permission, you would likely lose your case due to such copyright law.
One case that shows this issue is in the case, Andersen v. Stability AI Ltd where U.S. District Judge William Orrick indicated during a hearing in San Francisco that he was inclined towards dismissing the majority of lawsuits filed by artists against AI companies. This is in regards to the legal proceedings brought forth by artists Sarah Anderson, Kelly McKernan, and Karla Ortiz against Stability AI, DeviantArt, and Midjourney where the central focus revolves around the alleged unauthorized downloading of billions of copyrighted images by the Stable Diffusion software, created by Stability AI. This software, as claimed by the artists, was utilized as a “software library” to train various visual generative AI platforms, including DreamStudio, DreamUp, and Midjourney’s platform.
The court granted permission for specific copyright claims against Stability to move forward, but it recognized limitations in the artists’ claims against DeviantArt and Midjourney, citing unclear explanations in their theory on how Stable Diffusion utilized training images. Notably, the artists have been given the chance to revise their complaint (Dickstein and Delman).
And this is the current issue with copyright. Copyright was originally designed with human authors in mind — copyright law applies to works crafted by individuals. However, AI, being a tool, lacks the legal standing of a human creator.
Hence, these copyright laws are inadequate in protecting individuals whose art is pilfered by AI models. These shortcomings arise due to the challenge of providing conclusive evidence and the outdated nature of these copyright laws, enabling users to exploit AI-generated art freely as long as they remain undetected. Additionally, tracing back to a specific image within the vast datasets utilized by AI generators is nearly impossible — this results in the neglect of giving proper credit, compensating creators fairly, and showing due respect for the rights of others.
The reason why this all matters is because this shows a potential lack of safety for artists due to concerns about AI-generated art. Artists may feel apprehensive about sharing their work in an environment where AI art is prevalent, fearing some algorithm will immediately snatch their work and exploit their creations without proper attribution or consent.
This issue surrounding AI art is really reminiscent of current issues that go beyond art and copyright laws. This act of appropriating art resonates with cultural appropriation, where individuals from a majority group adopt elements from a minority group in an exploitative, disrespectful, or stereotypical manner, ultimately profiting from these actions.
A common instance of cultural appropriation would be within East Asian culture. Individuals often get tattoos of Chinese characters without understanding the language, leading to inaccurate translations and financial gains for the tattoo artists. Similarly, in the media, specific clothing styles or significant traditions may be ridiculed or sexualized for monetary purposes.
This issue is not confined to a specific region; it spans all cultures where exploitation persists due to the lack of effective laws that are implemented around these. This allows individuals to act disrespectfully without any consequences. Yet, traditional cultures get the short end of the stick as they remain vulnerable to appropriation for personal gain without sufficient laws in place to safeguard them (Vézina).
So when talking about art, this cultural appropriation and cultural invasion happens when artists find their artwork being appropriated by the wider public. Consequently, it acts as a deterrent for individuals considering pursuing majors in art.
For example, student Tyler Tran, who is an artist and student at Milpitas High School who was aspiring to major in digital art and animation in college, expressed deep concerns about the potential impact of the growing popularity of AI art on his future career prospects. He particularly emphasized the vulnerability of his future job opportunities, especially considering the significant cost associated with art education (Nguyen).
So without updated copyright laws to protect artists, the very essence of artistic expression is at stake when the legal frameworks designed for human authors prove insufficient for the challenges posed by AI-generated content.
In the end, it’s really about preserving the soul of art itself. This issue surrounding copyright leads us to challenge ourselves to define and defend the boundaries of creativity in a digital world. Creativity is a distinctly human trait, encompassing emotions, experiences, and nuances of existence that AI can not authentically ever replicate. Even the act of crafting something from one’s imagination hones skills and a unique perspective, defining an artist’s identity.
So while it may be near impossible to stop the spread of AI-generated art, this doesn’t mean that we should disregard the legal and logistical complexities surrounding such an issue. Rather, in essence, acknowledging the ethical dimensions of AI-generated art is not an attempt to stifle innovation but rather a commitment to uphold the values that define the artistic realm. It is a call to develop ethical guidelines, industry standards, and legal frameworks that protect the rights and dignity of human creators in this rapidly evolving landscape. Only then can AI-generated art even begin to enter the conversation on its ethical implications.
Works Cited
Dickstein, Tal, and Edward Delman. “Andersen v. Stability Ai Ltd..” Loeb & Loeb LLP, 30 Oct. 2023, www.loeb.com/en/insights/publications/2023/11/andersen-v-stability-ai-ltd.
Elgammal, Ahmed. “Ai Is Blurring the Definition of Artist.” American Scientist, vol. 107, no. 1, 2019, p. 18, https://doi.org/10.1511/2019.107.1.18.
Mathur, Atreya. “Art-Istic or Art-Ificial? Ownership and Copyright Concerns in AI-Generated Artwork.” Center for Art Law, 8 Sept. 2023, itsartlaw.org/2022/11/21/artistic-or-artificial-ai/.
U.S. Copyright Office. “What Is Copyright?”, U.S. Copyright Office, www.copyright.gov/what-is-copyright/.
Vézina, Brigitte. “Cultural Appropriation Keeps Happening Because Clear Laws Simply Don’t Exist.” Centre for International Governance Innovation, 24 Dec. 2019, www.cigionline.org/articles/cultural-appropriation-keeps-happening-because-clear-laws-simply-dont-exist/.
