In a recent video shared by French DJ David Guetta, a wave of apprehension rippled through the music community. Guetta expressed amazement at the success of an experiment that he initially considered a mere joke. In the clip, he enthralled a club full of fans with a track that seemed to sample rapper Eminem’s distinctive voice. However, the truth was far from what it appeared. Guetta had leveraged artificial intelligence (AI) technology to generate lyrics in Eminem’s style and then used another AI system to recreate the rapper’s voice. The result was a resounding success, leaving the crowd ecstatic.
The utilization of AI in music production has sparked growing concerns within the industry, with some insiders likening it to the disruptive impact of the file-sharing site Napster in the early 2000s. While the entry barrier for creating music was already relatively low, AI has further opened the floodgates, making it easier than ever to produce and distribute music. Websites like Boomy have empowered users to generate over 14 million songs, dwarfing Spotify’s entire catalog of approximately 100 million songs.
Lucian Grainge, the CEO of Universal Music, has raised alarms about the unchecked proliferation of generative AI. In a recent statement to investors, he emphasized the numerous dangers associated with this technology. Universal Music even sent a letter to leading streaming platforms cautioning them against allowing AI to train itself on copyrighted music, highlighting the pressing need to address copyright infringement concerns. While some artists, such as Grimes, are open to their voices being duplicated and royalties being split equally, the industry as a whole seeks to establish a framework for licensing music used by AI generators.
Universal Music’s worry extends beyond copyright infringement. The market share of major-label music on streaming platforms has been steadily declining, with the four largest suppliers’ portion dropping from 87% in 2017 to 75% in 2022. Independent artists, ambient tracks, and AI-generated songs are gaining popularity, diverting listeners away from major-label music. Grainge attributes this shift to an “oversupply” of content on Spotify, where a staggering 100,000 new tracks are added daily. He acknowledges that AI has played a significant role in this phenomenon.
This transformation has far-reaching implications not only for music companies but also for the future of music consumption. Spotify, once compared to Netflix, is now evolving into a hybrid platform akin to Netflix and YouTube. It provides access not only to professionally produced music but also to short audio clips like rain sounds, which can be effortlessly created by anyone with a computer.
AI has been a catalyst for this transformation, with some industry insiders describing AI-generated music as “user-generated content (UGC) on steroids,” reminiscent of the homemade videos, memes, and song covers dominating YouTube. Grainge and other industry leaders, including Warner Music CEO Robert Kyncl, are now contemplating an economic model for streaming that values professional music differently from user-generated content. Kyncl recently emphasized the need to differentiate between the value of an Ed Sheeran stream and the stream of rain falling on a roof.
What shape might this new economic model take? It is conceivable that user-generated music will find a separate platform, while professional music remains exclusive to premium services. However, Spotify might be reluctant to embrace such a partition. As the music industry faces further disruption, Mark Mulligan, an analyst at consultancy firm Midia, asserts that we have only witnessed the beginning of this paradigm shift.
The music industry is standing at the crossroads of innovation and uncertainty, where the interplay between AI, creativity, and economics will shape the future of music consumption.