YouTube Developing AI Tool Allowing Creators to Use Famous Musicians’ Voices

YouTube is currently in negotiations with record labels to launch a new AI tool that will allow creators to make videos using vocals from popular musicians, according to Billboard.

The trade publication reports that YouTube had hoped to introduce a beta of the tool with a small selection of artists to a “select group” of creators during its Made On YouTube event in September, but coming to licensing agreements with the major labels — Universal Music Group, Sony Music Entertainment, and Warner Music Group — delayed those plans (A wide launch would necessitate new agreements).

In addition, it has apparently been “challenging” to find big names to participate in the AI tool, with artists concerned about how creators could use their voices to deliver messages or sing lyrics that they don’t approve of.

Per Billboard’s sources, the labels view this deal with YouTube as a vital “framework” for the use of AI, with a few key points in the negotiations. In addition to the questions of how the AI model is trained and whether artists will have the option to either opt in (or out), figuring out monetization is a complicated task. While YouTube’s Content ID system is already able to identify and monetize copyrighted works in user-generated videos, the deal will have to sort out whether artists will get paid for their music being used as “an input” fed through the AI model or the “output” created by the tool.

And of course, figuring out how publishing rights factor into the equation is even more complex. YouTube would have to pay hundreds of songwriters whose work would be sampled for the beta tool and would likely prefer to pay a lump sum for licensing, which publishers would then have to divide among their writers.

Despite all these factors, label executives are invested in the idea of getting a deal done because they want to be seen as “proponents of progress.”  In April, the use of AI-generated soundalike vocals came to the forefront when an artist named ghostwriter released a song titled “heart on my sleeve” mimicking the voices of Drake and The Weeknd. After the track went viral, UMG issued a statement criticizing “infringing content created with generative AI.”

Since then, however, major labels have taken a softer stance as streaming services have acquiesced to requests to take down recordings using AI-generated vocals made to sound like popular musicians. Over the past several months, Warner Music CEO Robert Kyncl has acknowledged that the industry would have to “embrace the technology, because it’s not like you can put technology in a bottle,” while UMG CEO Lucian Grainge shared his belief that AI has the potential to “amplify human imagination and enrich musical creativity in extraordinary new ways.”

Meanwhile, musicians have been mixed in their reactions to the use of AI. Corey Taylor, Nick Cave, and Queen’s Brian May aren’t fans, but Paul McCartney and David Guetta have already embraced it in their work.