A viral song made with AI imitations of Drake and the Weeknd’s voices brought a difficult question into focus: who gets to use a musician’s voice, and under what terms? Some artists welcome AI as a creative tool, while others may discover their voices have been copied without their knowledge.
Artists are approaching AI in different ways
Musicians Holly Herndon and YACHT have used AI to explore new creative possibilities. YACHT trained an AI on 14 years of the group’s music and shaped the results into the album “Chain Tripping.” Herndon created Holly+, a website that lets people make deepfake music using her voice.
Those projects involve artists choosing to experiment with their own work or likeness. The viral Drake and Weeknd deepfake was different: neither artist took part in its creation. That gap between consent and unauthorized imitation is at the heart of the current debate.
Grimes has offered 50% royalties for AI-generated songs that use her voice, while also expressing interest in “killing copyright.” Her position highlights a tension around ownership: royalty arrangements depend on some way to recognize and compensate creators, while calls to dismantle copyright could undermine that system.
More music does not mean every artist benefits
On Spotify’s quarterly earnings call, CEO Daniel Ek said AI “should lead to more music” and that more music is “great culturally.” Spotify had taken down “Heart on my Sleeve,” the AI song using deepfakes of Drake and the Weeknd, but Ek sounded cautiously optimistic about the technology.
A larger supply of music may be good for a streaming business if listeners use its service more. For working musicians, however, the effects are less straightforward. Entertainment lawyer Henderson Cole told TechCrunch, “When artists are already struggling, it seems like a dangerous step.”
The source points to abysmal streaming payouts and COVID-19’s long-term impact on live music as pressures artists were already facing. If AI tools use a performer’s voice or work without consent, the question of who benefits becomes harder to ignore.
New tools can make the music industry a test case
Erickson described music as having a special social role in the development of technology. It can give emerging tools a use case, attract interest and help draw investment. That can put musicians in the position of testing technologies before the rules and norms around them are settled.
The source compares this pattern with crypto projects that promised to change music royalties and ticketing but had not reached anything close to mass adoption. It also points to sampling as an example of a technology and practice that did take hold. Sampling can be accepted when musicians get permission from the artist and label; cases without consent have led to high-profile lawsuits.
That history suggests that adoption alone does not settle questions of fairness. The way a new tool is used, and whether creators agree to that use, can shape how it fits into music culture.
Copyright questions remain unsettled
Some copyrighted material can be used without explicit permission under fair use. The analysis considers whether a work was made for profit, how much copyrighted material it uses, how transformative it is and whether it could harm the original’s market.
Cole believes a fair use argument for AI music would be unlikely to carry much weight in practice. He pointed to lawsuits involving Ed Sheeran and Robin Thicke over alleged similarity to hit songs, and said that allowing AI to copy an artist’s voice or musical sound seems unlikely.
Legal systems take time to respond to new technology, while major labels are already stating their position. A Universal Music Group representative said the company has “a moral and commercial responsibility” to prevent unauthorized use of artists’ music and stop platforms from ingesting content that violates creators’ rights.
The representative framed the debate as a choice between supporting artists, fans and human creative expression, or enabling deepfakes, fraud and denying artists compensation. As AI music develops, consent and payment remain central questions for the people making, distributing and listening to it.