As the music industry wrestles with how to incorporate, and value, AI work, a third of artists are already using it for some form of inspiration.
A new study from the Boston’s prestigious Berklee College of Music found that 33 percent of respondents “use AI to generate initial ideas, melodies, or reference tracks that are later reworked.” Perhaps even more surprisingly, a number nearly as large put it in the final product — about 26 percent of artists “use AI for full backing tracks in finished work.”
“Had we done this study 18 months ago, these numbers would have been a lot lower,” Mark Ethier, who helped run the research, told The Hollywood Reporter, “and from that perspective it’s pretty shocking.”
A somewhat smaller percentage, 18 percent, will play with the tools in the creation process but keep them out of the released track.
But in contrast to conventional wisdom newer musicians were less likely to use AI: artists who were “just starting out” used AI in “any way” at a rate of 56 percent while “full time” creators did so all the way to the tune of 92 percent.
The study was conducted by the Berklee Emerging Artistic Technology Lab (BEATL), where Ethier serves as executive director, as part of the AIMS: AI Music Summit the school just held on its campus Wednesday-Friday and focused on AI music in video, as opposed to artists recording only sounds; the numbers of course could be different in that population.
The Berklee study was conducted among more than 1,000 people associated with the music business, from artists to marketers, music supervisors and video creators.
Jim Lucchese, Berklee’s president, told The Hollywood Reporter that he saw the numbers as evidence not only of sharply divided opinions on AI among artists but different attitudes even within the group of those embracing the tech.
“We’re seeing a high level of individual creative agency in how differently AI can be used, from a composition tool to just pushing your skills.”
The study comes at a time when the music industry is grappling with how to use AI generations in the work and how artists should be compensated when they do. On Friday, the American Federation of Musicians sued Universal Music Group and Warner Music Group over a settlement the giants made with AI music firms Suno and Udio, alleging that the labels “have refused to compensate the musicians whose work … is fed into AI machines for profit.”
In keeping with the significant minority using it for initial ideas, 31 percent of Berklee respondents use AI for “lyric generation,” presumably in finished work.
Those who do incorporate AI into their music may be limiting their potential audience, however. A THR/Frost School of Music poll from the fall found that 52 percent of respondents were “not interested” in listening to music from their favorite artists made with the help of AI.
In addition to established artists sprinkling AI into their work, the past year has seen hits from entirely synthetic acts that involved no playing of instruments, from Breaking Rust’s country smash “Walk My Walk” to Saxboy Billy’s electropop TikTok sensation “The Puerto Rico Song.”
The music business has been particularly wracked by AI changes given the industry’s long history of incorporating new technologies along with the fact that an AI generation can be harder to spot than in a video, which for now can still sometimes descend into the uncanny valley.
Supporters argue technology and sampling have long been a part of the soundscape; opponents say releasing music derived, fundamentally, from what someone else wrote is neither original nor ethical.
As has been the case with podcasts, video is playing an increasing role in launching a successful music career. More than 75 percent of those Berklee asked said they believe it “directly shapes career outcomes.”
Leave a Reply