While copyright in the U.S. was formerly instituted by Act of Congress in 1790, its origins lie in several hundred years of prior legal evolution driven by the introduction of the printing press. Ever since, copyright has been in a dance with technology, as new forms of expression made possible by new inventions push the law to consider and adapt.
More recently, the boundaries of which works may qualify for a copyright were tested in 2011, when a macaque named Naruto took a selfie using a camera deliberately positioned by a wildlife photographer in the jungle. In the ensuing litigation, the question followed: who owned the copyright to the selfie? The photographer; the monkey; or no one? More recently in the 2020s, large language model AI programs trained on copyrighted material scraped from the internet have raised even more questions, including whether the underlying copyrights were infringed? And for the work generated by the program, who owns it? The holders of the original copyrighted materials from which the AI generated its content; the AI company; the person requesting the AI work; no one at all; or all of the above?
AI in particular is testing the boundaries of creative industries at an unprecedented and not always welcome pace, at times challenging traditional applications of copyright law. While AI’s ability to generate text, art, and music offers exciting possibilities, it also raises thorny legal questions. Who owns the rights to AI-generated content? What happens when an AI system is trained on copyrighted works? And what ought law makers, government agencies, and courts do in balancing various competing interests?
Copyright’s Long Dance with Technology
Copyright law is in no small part a product of technological innovation. The invention of the printing press, photography, film, sound recording, digitalization and the rise of the internet have forced legal frameworks to adapt to new forms of expression and distribution. For instance, when photography first emerged, courts had to decide whether a photograph was a mere mechanical reproduction or a creative work deserving copyright protection. Similar debates resurfaced with the advent of film and digital media. In each case, copyright law was tested, and in discrete ways expanded and redefined to accommodate new technologies and their products.
Authorship, Originality, and Training Data
At its simplest, copyright is a type of intellectual property that protects an original work of authorship once its author fixes the work in a tangible form of expression. To be original, the work must be independently created by a human author, and demonstrate at least a minimal degree of creativity–without copying. While the tangible form in which a work is made may change, the human and creative elements cannot. But how attenuated can the relationship between the work and its author become?
In the case of Naruto, the Ninth Circuit Court of Appeals decided that as a monkey, he could not hold a copyright to the selfie. Whether the selfie was itself a copyrightable work in terms of the photographer that made it possible has remained more controversial, although in 2014, the U.S. Copyright Office affirmed that works created by non-humans are not copyrightable.
Further building on the “human author” requirement, in 2023, the U.S. Copyright Office clarified its stance on AI generated content, that because it lacks meaningful human input, it is not eligible for copyright protection. This decision emphasizes the importance of human creativity in the copyright framework. However, it also leaves creators in a grey area when they use AI tools as part of their creative process. If a photographer uses AI to enhance an image or an artist relies on AI-generated elements, who owns the final product? These are the questions courts and lawmakers will need to address in the coming years.
Another flashpoint concerns the training of AI models. Generative AI relies on vast datasets, often compiled from copyrighted works, to learn how to produce new content. This practice has raised alarms about potential copyright infringement. Are AI developers violating copyright law when they feed AI systems protected works without permission? Some argue that using copyrighted material for training falls under “fair use,” a doctrine that allows for limited uses of copyrighted works without permission. However, this argument has yet to be fully tested in court, leaving the legal status of AI training data unresolved.
Public Input and the Path Forward
In response to these emerging challenges, the U.S. Copyright Office has taken a proactive approach. In 2023, it issued a Notice of Inquiry and hosted listening sessions to gather input from various stakeholders, including creators, tech companies, and legal experts. The goal was to explore how copyright law should evolve to address the implications of AI.
One of the key areas under examination is the creation of digital replicas, such as AI-generated likenesses of real people. The potential for AI to mimic a person’s voice or image without their consent has sparked concerns about privacy, publicity rights, and copyright law. As AI-generated content becomes more realistic and widespread, it is more likely to prompt new legislation that seeks to balance the rights of individuals and copyright holders with the technological possibilities AI brings. Indeed, one early entrant into that fray is Tennessee, which passed the ELVIS Act, establishing strong protections against unauthorized deep-fakes and voice clones.
Shaping the Future of Copyright and AI
Balancing the protection of human creativity with the need for technological innovation is no easy task. But one thing is clear: AI is changing how we think about creation itself. As copyright law adapts to this new reality, businesses will need to protect their IPs by striking a delicate balance between rewarding human ingenuity and embracing the possibilities of machine-generated innovation.
At Baker Jenner we understand your company’s need to remain competitive in the great AI race. In the meantime, keep in touch with us regarding the evolving copyright laws. Contact us today at (404) 400-5955 or here to discuss what we can do for you.
Baker Jenner LLLP
