By: Lakshmi Sanmuganathan
On April 4, 2023, the track titled “Heart on My Sleeve” was uploaded to social media and streaming services by TikTok user Ghostwriter977. Generated by artificial intelligence (AI), the song garnered rapid, viral attention for its uncanny ability to mimic the singing, lyrical, and production style of Canadian musicians Drake and The Weeknd. Within days, the original video, which amassed over 9 million views, was removed from online platforms in response to copyright violation claims by the artists’ record label, Universal Music Group. This was the latest example of one of 2023’s most controversial trends – the use of copyrighted material to create AI-generated art.
There is a rising legal tension between the promotion of AI innovation and the protection of human creativity. Amid this unfolding conflict, governments and international organizations have offered divergent responses to this ongoing debate.
On August 3, 2023, the United Kingdom published a white paper, “A Pro-Innovation Approach to AI Regulation,” which emphasized the state’s “goal to become a science and technology superpower by 2030,” through the promotion of AI technology innovation. The policy document sets forth a “pro-innovation regulatory framework,” which emphasizes the UK’s newfound freedom following its exit from the European Union (EU) to craft regulations that will position the UK to become an AI powerhouse. The commitment to a pro-innovation approach is not only a strategic move to establish the UK as an AI leader, but also serves to incentivize overseas businesses to establish a presence in the UK amid growing regulatory pushback in other countries. While the white paper noticeably mentions the risk of AI to mental health, privacy, and human rights, the policy proposal fails to address intellectual property (IP) concerns, illuminating the UK’s increasing deference towards AI innovation in the face of human creation concerns.
The UK’s emerging pro-innovation stance arrives at a time when governments around the world and international organizations, such as the World Intellectual Property Organization, are grappling with how to adapt legal frameworks to consider AI advancements. Japan and Singapore have already passed legislation that amends their copyright frameworks, permitting lawful access to copyrighted works for the purpose of machine learning and AI innovation. These emerging copyright exceptions raise concern for creative sectors worldwide, including the music, literature, film, and photography industries.
Like the European Union, the United States is not presently confronted with legislative proposals to allow AI copyright exceptions. However, the overall state of U.S. copyright law in the face of AI advancements remains chaotic. The US Digital Millennium Copyright Act created a process that allows copyright holders to notify third parties, such as Spotify or YouTube, about alleged copyright violations in their distribution process. Such notifications create obligations for third-party distributors to swiftly takedown materials in a process traditionally known as a “DMCA takedown.” However, creators of AI-generated music argue that their work is not a copyright infringement, citing protections under the First Amendment’s freedom of speech and fair use doctrines.
While copyright grants exclusive rights to authors, U.S. copyright law supplies a number of limitations and exceptions, with the most significant being the application of the fair use doctrine. Codified in the 1976 Copyright Act, the fair use doctrine permits users to use copyrighted works without permission, under specific circumstances and for various purposes, such as criticism, comment, teaching, research, and scholarship. Justice Ginsburg referred to the fair use doctrine as a “built-in free speech safeguard,” which ensures that copyright serves the public good. The doctrine prevents copyright holders from possessing unlimited freedom to control all uses of their works, particularly when faced with public interest considerations related to transformative production.
Advocates for AI-generated music apply the fair use doctrine to argue that AI training is not conducted to copy commercial products but is instead intentionally transformative. In pursuit of innovation, proponents of AI-generated music contend that allowing an AI machine to access copyrighted materials serves the primary purpose of training an algorithm, with no direct impact on the market of the original product.
Amidst competing copyright safeguards and constraints, artists have turned to right of publicity arguments in the United States, which have demonstrated a higher level of predictive success. In 1988, the Ninth Circuit held that the intentional imitation of a professional singer’s distinctive voice to sell something is a violation of the right of publicity. This provides strong legal support for artists, such as Drake and The Weeknd, who are battling to protect their likenesses against AI-generated music. However, there is no current federal right to publicity in the U.S. and only 14 states have specific statutes covering such a right. Additionally, publicity protections vary from state to state. While most states require brands to secure written permission from an individual to use their name, portrait, picture, or voice for advertising or commercial purposes, rules vary across states in relation to parodies, digital recreations, and posthumous rights.
Which deserves greater protection: human or machine creation? The evolving landscape of United States copyright law in response to AI innovation marks a legal battle in its infancy that remains to be settled in the legislative and litigation arenas. The outcome of such conflicts will offer essential clarity on how the law should balance the interests of human creativity with the temptations of technological advancement.
