Post Authored By: Natalie Elizaroff
Another viral trend is sweeping social media – and this one has become a real legal doozy. Artificial Intelligence (AI) art has generated a wave of controversy and unanswered questions as magical avatars and AI-created landscapes flood the internet. Not only are lawyers left scratching their heads for answers to never-before-seen problems, but artists are in a frenzy to get their voices heard as the situation barrels out of control.
Understanding AI Software in the Limited Scope of Artistry
AI is a complex and often misunderstood concept that has continued to grow and expand as humanity evolves. To put it simply, AI is defined as “a field that combines computer science and robust datasets to enable problem-solving.” [1] To narrow this definition even further, in the context of artistic ventures, AI art refers to art generated with the assistance of artificial intelligence. AI learns and operates through hundreds of thousands of algorithms that derive its knowledge from datasets. Accordingly, the creative artworks resulting from AI heavily rely on the quantity, quality, and type of data the AI is fed.[2]
Controversies and issues have arisen because to create the astonishing work that AI has produced due to various prompts, the software must be fed real artists’ work in order to grow the dataset and expand the AI knowledge base. AI applications like DALL-E 2, Midjourney, and others, are built by scraping millions of images from the open web, then teaching algorithms to recognize patterns and relationships in those images and generate new ones in the same style. [3]Accordingly, artists uploading their work onto the internet to places like DeviantArt, ArtStation, Tumblr, Pinterest, and more may be unwittingly helping train the AI algorithms that are mimicking their work. It is a vicious cycle that has many artists seething, not only from the injustice and lack of credit but from how information and understanding of the software is being mishandled.
An Artist’s Perspective
To get a better grasp of how this AI dilemma is affecting artists and creators, I sat down with an artist that goes by the pseudo name ‘KrowMagg’. KrowMagg has been involved in art for many decades, going back to their first illustration commission in high school that ended up being featured in a publication. KrowMagg has a degree in Multimedia Design with a focus on game art and animation, so they are familiar with digital art creation. They have worn a variety of hats and have spent considerable time working in packaging design and graphic design, as well as more complex industrial and electromechanical design. When asked to share their thoughts on the situation at hand, they had the following to say:
History of art culture has been evolving from the dawn of humanity to capture the imagination of the mind and imagery of the world around them. Though, the reproduction of imagination and worldly imagery has been a labor of chasing the mist where the “talented few” could capture an approximation of what they “see.” The evolving chase has given us new tools and technology to “capture” the mist of imagination and imagery from charcoal etchings to bits and bytes projected on screens. In this digital age, great imagery is the key to success and money.
People are impatient, time is money, and people don’t value artistic abilities. Enter “Fiverr.” Get a quick image and slap it on the product and get it to market fast. The downside to gig jobs is the most experienced artist wins… to be the most experienced, the artist must train for years for their artistic expression, and training takes time. Enter “AI.” If the artist costs too much and takes too long to create great images, they train computers to do it faster. The thing is, they use artist work to train AI. The same artist that trained for years with a huge backlog can be on demand at the drop of a dime with the help of AI. Just looking at it and you can get a thousand iterations of a “this is fine” meme in the “style” of Norman Rockwell. This is the part that bothers me. The “style” is only derived from using Norman Rockwell’s work to train the AI. The Algorithm just mimics the stroke, style, and emotion of the reference, and people call it original art and even sell the AI renders at conventions.
I know art is built on each other. We learn from the masters, but how would the record industry respond to an AI rendition of a Metallica-style free music library where the AI is taught by ripping original songs from Metallica to then rebuild the “new” music sounding just like the original band?[4]
On the other side of the table, some people are embracing the AI influx of art and even profiting from it. Jason M. Allen, founder and lead developer of board game company Incarnate Games, recently won the “digital arts/digitally-manipulated photography” category at the Colorado State Fair. Mr. Allen used Midjourney, an AI software that turns lines of text into graphic imagery, to generate his “Théâtre D’opéra Spatial” submission. Mr. Allen submit the piece under the name “Jason M. Allen via Midjourney”, which satisfied the fair’s rules allowing any “artistic practice that uses digital technology as part of the creative or presentation process.”[5]
The waters get even murkier considering that Kris Kashtanova, a designer and AI enthusiast, earlier this year announced that they received the first known U.S. copyright registration for artworks created by an AI-image generator (Midjourney).[6] Ms. Kashtanova followed up by explaining their use of AI-image generators was limited as a tool to assist the overall creation of her work. She explained that her comic book, Zarya of the Dawn, was not entirely made by AI and that Ms. Kashtanova wrote the comic book story, designed the layout, and made other artistic choices. In a surprising turn of events, likely fueled by the frustration and anger of the art community, the Copyright Office recently reversed its decision and announced that the copyright protection on the comic book will be revoked.[7] Only a few short months after granting Ms. Kashtanova the copyright, the Copyright Office explained that copyrighted works must be created by humans to gain official copyright protection and that it errantly missed Midjourney’s involvement.
The Art Community’s Response
Although someone like Mr. Allen is reveling in his winning AI-generated masterpiece, a majority of the art community is outraged. As the AI-generated Art trend has continued to grow, artists have spoken out and begun including “No AI Art” images in their portfolios.[8] The consensus among the masses is that although AI is a neat tool, it is disrespectful to feature its text-prompt creations as a mainstay or alongside artist-rendered artwork that took hundreds of hours to create. Some online art communities have responded, such as ArtStation, by stating that although AI-generated artwork on the site will not be banned, it will incorporate tags “enabling artists to choose to explicitly allow or disallow the use of their art for (1) training non-commercial AI research, and (2) training commercial AI.”[9] Others, such as Getty Images and Shutterstock, have outright banned AI-generated submissions due to the ‘legal uncertainty’ that surrounds AI images.[10] Whereas some AI software developers have responded by providing a way for artists to opt out of training their software. This solution received mixed reviews given that the default state would still be ‘opted in’ and unless someone was aware of the option, their artwork is automatically being included and fed to the AI.
The Law is Obscure & AI-Artists Should be Wary
The legal field is of no help to this ongoing dilemma and existing laws are divided depending on the country. The United States Copyright Office provides a clear declaration by stating that it will “register an original work of authorship, provided that the work was created by a human being.”[11] This is consistent with US case law specifying that copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind.”[12] Despite this clear explanation, situations like Ms. Kashtanova’s make copyright and AI-generated art difficult to balance.
Likewise, many artists are wondering whether they have any legal grounds to prevent AI artists from claiming ownership of art that is directly derived from their work. This is often where the ‘fair use’ argument crops up and individuals rely on whether the resulting artwork is sufficiently transformative. Courts have not yet ruled on any substantive cases that discuss this dilemma, but as AI artwork becomes more prominent there is no question that this issue will soon arise. Despite the lack of established caselaw, AI artists should be wary and make sure to document their creative process of text-prompt generation. Fair use is a reasonable argument to employ, but if the resulting artwork is substantially similar to its underlying work, it will likely fail the transformative analysis.
Final Thoughts
Artificial intelligence is not going away. To proactively approach this problem, lawmakers must determine what kind of protection should be granted to intelligent algorithms, if any. Although individuals like Mr. Allen may believe that “Art is dead…It’s over. A.I. won. Humans lost,” we still have a chance to adjust the algorithm and change the trajectory.
[1] IBM Cloud Education, Artificial Intelligence (AI), International Business Machines(June 3, 2020), https://www.ibm.com/cloud/learn/what-is-artificial-intelligence.
[2] Adam Hencz, Agents Of Change: Artificial Intelligence – AI Art and How Machines Have Expanded Human Creativity,Artland Magazine (April 8, 2022), https://magazine.artland.com/ai-art/.
[3] Kevin Roose, An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy,N.Y. Times (Sept. 2, 2022),https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html.
[4] To learn more about KrowMagg and to support their art, please visit: https://krowmagg.art/
[5] Id.
[6] Matt Growcoot, AI-Generated Artwork is Copyrighted for the First Time, PetaPixel (Sept. 27, 2022), https://petapixel.com/2022/09/27/ai-generated-artwork-is-copyrighted-for-the-first-time/.
[7] Brian Cronin, AI-Created Comic Has Been Deemed Ineligible for Copyright Protection, Comic Book Resources (Dec. 20, 2022), https://www.cbr.com/ai-comic-deemed-ineligible-copyright-protection/.
[8] Benj Edwards, Artists stage mass protest against AI-generated artwork on ArtStation, ARS Technica (Dec. 15, 2022), https://arstechnica.com/information-technology/2022/12/artstation-artists-stage-mass-protest-against-ai-generated-artwork/.
[9] Id.
[10] Harrison Jacobs, Getty Images Bans AI-Generated Images Due To Copyright Worries, Art News (Sept. 22, 2022), https://www.artnews.com/art-news/news/getty-images-bans-ai-generated-images-due-to-copyright-1234640201/.
[11] U.S. Copyright Office, Copyrightable Authorship: What Can Be Registered (Jan. 28, 2021), https://copyright.gov/comp3/chap300/ch300-copyrightable-authorship.pdf.
[12] Feist Publ’ns v. Rural Telephone Serv. Co., Inc.,499 U.S. 340 (1991).
About the Author:
Before pursuing a legal career, Natalie spent several years in the microbiology department at Evanston Hospital where she conducted comparative research studies, performed quality control testing, and worked on state-of-the-art medical device technology. After doing a swift 180 and finding law as her true calling, Natalie focused her efforts into intellectual property.
Natalie received a Bachelor of Science in Molecular Biology, with a minor in Biostatistics from Loyola University Chicago. She earned her law degree from UIC School of Law and she is currently working as an Associate at Advitam IP LLC, where she handles a variety of IP matters including trademark litigation, copyright infringement, and other IP-related disputes.