In April 2023, a song called “Heart on My Sleeve,” created by a mysterious producer known as Ghostwriter, gained significant popularity on TikTok and momentarily became the top song on both YouTube and Spotify. However, it was quickly removed from these platforms. The producer used artificial intelligence to create vocals that sounded like Drake and The Weeknd, leading Universal Music Group, which represents these artists, to threaten legal action. Despite being aware of the situation, Drake didn’t seem troubled. In fact, he later used AI-generated vocals in his music during his ongoing rivalry with rapper Kendrick Lamar.
I’ve been closely monitoring these events as a scholar of digital media and a rap artist who was among the first to mix rap lyrics with samples of previously released vocals. AI can help artists produce music, as Drake demonstrated in his diss track. However, the technology exists in a legal gray area, especially concerning vocals. On April 19, 2024, Drake released a song called “Taylor Made Freestyle,” which featured AI-generated vocals mimicking Tupac Shakur and Snoop Dogg. The AI voice of Shakur – who died in 1996 – directs sharp criticism at Lamar, taunting him for his silence in the feud between the two rap legends.
Shakur’s estate unsurprisingly threatened legal action against Drake for using Tupac’s voice and persona without authorization, claiming it violated the late artist’s rights to control the commercial use of his identity. Howard King, the estate’s attorney, emphasized that the estate would never have approved this usage. Consequently, Drake removed the diss track from streaming platforms and YouTube.
It’s crucial to distinguish between copyright and someone’s right to publicity. Copyright law traditionally applies to works authored by humans, meaning only humans can be considered authors, and their work cannot be used without their permission. One key legal issue with AI and copyright is the use of copyrighted material to train AI models, exemplified by The New York Times suing OpenAI and Microsoft for training their models on the publication’s articles without permission.
The right to publicity, in contrast, allows individuals to profit from their name, image, likeness, voice, or signature. One of the most notable cases involved Bette Midler suing Ford Motor Co. in 1988 for using a former backup singer to impersonate her voice in an ad, resulting in a $400,000 judgment in her favor. This case set a precedent that could significantly impact how AI-generated voices of celebrities are handled legally.
Scarlett Johansson might face a complex legal battle if she sues OpenAI for an AI voice assistant that sounds like her, despite the company’s claim that there was no intent to replicate her voice. Currently, federal copyright law doesn’t specifically address cloned vocals or the use of someone’s voice in a new context.
Although cloned vocals often feature original lyrics and musical elements provided by musicians prompting the AI, making them somewhat distinct from existing copyrighted materials, California and other states have set precedents asserting that impersonating a famous musician can violate their right of publicity. Such rights protect a person’s likeness, including their voice, even in new contexts. Historic cases, such as injunctions against musicians using Michael Buffer’s famous “Let’s Get Ready to Rumble!” catchphrase, demonstrate this protection, though these rights were mainly invoked in advertisements and commercial uses.
Given the legal ambiguity, the recording industry and other creative sectors are advocating for new legislation to address the issue. Recently, Tennessee passed the ELVIS Act, which extends publicity rights beyond advertisements to include voice cloning. Federal lawmakers are also considering similar legislation to create broader definitions of publicity rights.
In the wake of AI advancements, it is widely recognized that protecting the role of humans in creating art is essential. While AI can produce impressive imitations, it lacks the soul and spontaneity of human artists. AI should enhance and support artists’ work rather than overshadow it.