I got an email last October that had a heck of a beginning collection:” I fired a bomb at the US Copyright Office this morning. ”
The communication was from Elisa Shupe, a 60-year-old withdrew US Army veteran who had just filed a trademark registration for a book she’d just self-published. She’d used OpenAI’s ChatGPT heavily while writing the book. Her application sought to convince the US Copyright Office to reverse its scheme, which typically requires would-be rights holders to forbid the use of machine-generated parts in work produced by AI.
That original picture did n’t detonate—a year after, the USCO rejected Shupe’s application—but she finally won out. After Shupe appealed, the company changed its mind about publishing her autofiction book AI Machinations: Tangled Webs and Printed Words, which was self-published on Amazon under the pseudonym Ellen Rae, earlier this month.
The tale draws from Shupe’s adventurous lifestyle, including her advocacy for more equitable gender recognition. As more people incorporate AI devices into their creative work, its register provides a picture of how the USCO is dealing with artificial intelligence. It is one of the initial creative works to be granted rights for the layout of AI-generated text.
“We’re seeing the Copyright Office struggling with where to draw the line, ” intellectual property attorney Erica Van Loon, a companion at Nixon Peabody, says. Because of a major caveat required to be granted, Shupe’s case highlights some of the complexities of that struggle.
The USCO see authorizing Shupe to use her book for the first time as creator of the entire text is not acknowledged as is customary for written works. Rather she is considered the creator of the “selection, cooperation, and arrangement of text generated by artificial intelligence. This means that no one can backup the book without getting permission, but the real sentences and paragraphs are not and could potentially be rearranged and republished as a unique book.
The agency’s original attempt to file her work was October 10, the year of the agency’s trademark registration. It made no comment on this account. According to Nora Scheland, a representative for the firm, the Copyright Office does not comment on certain copyright licenses or pending registration applications. The company was asked to make recommendations to the White House regarding the “scope of shelter for functions produced using AI” in President Biden’s executive order on AI next fall. ”
Shupe actually requested that the USCO establish a more significant way to copyright recognition for AI-generated content, despite her distinctive limited copyright registration. “ I seek to copyright the AI-assisted and AI-generated material under an ADA exemption for my many disabilities, ” she wrote in her original copyright application.
Shupe passionately thinks that she was only able to finish her book using relational AI techniques. She claims that the Department of Veterans Affairs has classified her as 100 percent crippled and that she struggles to read because of cognitive impairment caused by conditions like borderline personality disorder, bipolar disorder, and brain stem deformity.
She is pleased with the finished product and considers using a language machine to express her thoughts in a different but still valuable way. “You do n’t just hit ‘generate’ and get something worthy of publication. That may come in the future, but we’re still far from it, ” she says, noting that she spent upwards of 14 hours a day working on her review.
After her first subscription was turned down, Shupe got in touch with Jonathan Askin, the leader of the Brooklyn Law Incubator and Policy Clinic at Brooklyn Law School, who handles pro bono cases involving emerging technology and policy issues. Sofia Vescovo, a student at Askin and Brooklyn Law, started working on Shupe’s event in January and appealed the decision to the USCO.
Shupe used ChatGPT as assistive technology to communicate, and the appeal drew on her explanation that her disabilities should be protected by copyright. She also compared her use of OpenAI’s robot to an armless using a prosthetic leg. The USCO alleged in the charm that she was discriminated against because of her illness. ”
Additionally, according to the Brooklyn Law elegance, Shupe should be able to obtain rights for compiling the book, or for organizing and selecting the text that has been generated by AI. It kept a detailed log of Shupe prompting ChatGPT, including personalized commands and edits she made.
A side-by-side assessment of the final version of Shupe’s publication and the machine output is included. On a word levels, she adjusted almost every line in some way, from shifts in word choice to construction. One way to describe a character in the novel is, “Mark eyed her, a complex combination of concern and annoyance evident in his gaze, ” which turns out to be, “Mark studied her, his gaze reflecting both worry and irritation. ”
The appeal makes reference to another recent AI copyright ruling involving the graphic novel Zarya and the Dawn, which incorporates AI-generated images created with Midjourney. Even though the authors were denied copyright on the specific images themselves, author Kris Kashtanova was granted copyright for the selection and arrangement of AI-generated images in the text in February 2023.
When the USCO granted Shupe’s request for copyright, it did not address the disability argument put forth but agreed with the appeal’s other argument. Shupe could be considered the author of “selection, coordination, and arrangement of text generated by artificial intelligence, ” the agency wrote, backdating her copyright registration to October 10, 2023, the day that Shupe had originally attempted to register her work. That overall confers her authorship of the work, preventing unauthorized wholecloth reproduction of the entire book but not copyright protection over the book’s actual sentences.
“Overall, we are extremely satisfied, ” says Vescovo. The team believed that copyrighting the book’s compilation would give readers peace of mind when they were forced to copy the work completely. We really wanted to make sure we could immediately provide her with this protection. The Brooklyn Law team hopes that Shupe’s approval can serve as a guideline for others who are experimenting with AI text generation and want some copyright protection.
“I’m going to take this as a win for now, ” Shupe says, even though she knows that “in some ways, it ’s a compromise. She claims that she should be able to copyright the actual text of the book and that the way she uses ChatGPT more closely resembles a collaboration than an automated output.
What the USCO granted Shupepe is known as” thin copyright,” a professor of law and artificial intelligence at Emory University, and that does n’t stop someone from rearranging the paragraphs into a different story. This kind of copyright, according to Sag, is what you would expect from an an anthology of poetry without writing it.
Erica Van Loon agrees. “It’s hard to imagine something more narrow, ” she says.
Shupe is a part of a larger effort to make copyright laws more welcoming to AI and its users. The Copyright Office, which oversees the copyright registration system and provides advice to Congress, the judiciary, and other government agencies on copyright issues, is a key player in shaping how works that use AI are handled.
The USCO has shown openness to registering works with AI elements despite still defining authorship as an entirely human endeavor. More than 100 works with AI incorporated have been registered, according to the USCO in February; a search by WIRED found over 200 copyright registration applications explicitly disclosing AI elements, including books, songs, and visual artworks.
One such application came from Tyler Partin, who works for a chemical manufacturer. He recently registered a tongue-in-cheek song he created about a coworker, but excluded lyrics that he spun up using ChatGPT from his registration. Partin views the text generator as a tool, but he does n’t believe he deserves to take credit for its output in the end. Instead, he applied only for the music rather than the accompanying words. “ I did n’t do that work, ” he says.
But there are others who share Shupe’s perspective and agree with her mission, and believe that AI-generated materials should be registrable. Some high-profile attempts to register AI-generated artworks have resulted in USCO refusals, like artist Matthew Allen’s effort to get his award-winning artwork Théâtre D’opéra Spatial copyrighted last year. The AI system he created deserves copyright protections of its own, claims researcher Stephen Thaler, who has spent years trying to demonstrate this.
Thaler is currently challenging a US court ruling from last year that rejected his request for copyright for his computer. The Artificial Inventor Project, a group of intellectual property attorneys who file test cases requesting legal protections for AI-generated works, was founded by Ryan Abbott, the lead attorney on the case.
Abbott is a supporter of Shupe’s mission, although he’s not a member of her legal team. He is unsatisfied that the copyright registration does not cover the AI-generated work itself. “We all see it as a very big problem, ” he says.
Although it is not a settled issue, Shupe and her legal team do n’t intend to further the ADA argument by challenging the USCO decision. Askin says that lobbying Congress for an ADA statute is probably the best course of action. We might write some legislation or provide testimony to try to move Congress in that direction. ”
In the age of AI, Schuppe’s qualified victory is still a significant indicator of how the Copyright Office struggles with what it means to be an author. She hopes that making her efforts known will lessen the stigma she perceives as a barrier to using AI as a creative tool. Although her metaphorical nuke did n’t explode, she has still furthered her cause. “ I have n’t been this excited since I unboxed a Commodore 64 back in the 1980s and, after a lot of noise, connected to a distant computer, ” she says.
