YouTube Creators Can Now Use AI to Imitate Pop Stars


Demi Lovato, Troye Sivan, Charli XCX and more are participating in a new experiment that will generate audio snippets for YouTube Shorts.

Demi Lovato

Angelo Kritikos*

YouTube is launching an experimental feature Thursday (Nov. 16) that will create artificial intelligence-generated voices of well-known artists for use in clips on YouTube shorts. The initial selection of acts participating in the program includes Charlie Puth, John Legend, Sia, T-Pain, Demi Lovato, Troye Sivan, Charli XCX, Alec Benjamin and Papoose. 

YouTube’s feature, called Dream Track, creates pieces of music — voice along with musical accompaniment — based on text prompts that are up to 30 seconds in length. For now, around 100 U.S.-based creators will have Dream Track access.

“At this initial phase, the experiment is designed to help explore how the technology could be used to create deeper connections between artists and creators, and ultimately, their fans,” according to a blog post from Lyor Cohen, global head of music, and Toni Reid, vp of emerging experiences and community.

The music industry has been wary of AI this year, but several prominent executives voiced their support for Dream Track. “In this dynamic and rapidly evolving market, artists gain most when together we engage with our technology partners to work towards an environment in which responsible AI can take root and grow,” Universal Music Group chairman and CEO Lucian Grainge said in a statement. “Only with active, constructive and deep engagement can we build a mutually successful future together.”

“YouTube is taking a collaborative approach with this Beta,” Robert Kyncl, CEO of Warner Music Group, said in a statement of his own. “These artists are being offered the choice to lean in, and we’re pleased to experiment and find out what the creators come up with.” 

YouTube emphasized that Dream Track is an experiment. The artists involved are “excited to help us shape the future,” Cohen said in an interview. “Being part of this experiment allows them to do it.” That also means that, for now, some of the underlying details — how is the AI tech trained? how might this feature be monetized at scale? — remain fuzzy.

While the lawyers figure all that out, the artists involved in Dream Track sounded enthusiastic. Demi Lovato: “I am open minded and hopeful that this experiment with Google and YouTube will be a positive and enlightening experience.” John Legend: “I am happy to have a seat at the table, and I look forward to seeing what the creators dream up during this period.” Sia: “I can’t wait to hear what kinds of recipes all you creators out there come up with.” 

While YouTube’s AI-generated voices are likely to get the most attention, the platform also announced the release of new AI music tools. These build on lessons learned from the “AI Music Incubator” the platform announced in August, according to Demis Hassabis, CEO of Google Deepmind. Through that program, “some of the world’s most famous musicians have given feedback on what they would like to see, and we’ve been inspired by that to build out the technology and the tools in certain ways so that it would be useful for them,” Hassabis explained in an interview.

He ticked off a handful of examples: An artist can hum something and AI-powered technology will create an instrumental based on the tune; a songwriter can pen two musical phrases on their own and rely on the tools to help craft a transition between them; a singer can come in with a fully fledged vocal melody and ask the tech to come up with musical accompaniment.   

Finally, YouTube is rolling out another feature called SynthID, which will watermark any of the AI-generated audio it produces so it can be identified as such. Earlier this week, the platform announced that it would provide labels and others music rights holders the ability “to request the removal of AI-generated music content that mimics an artist’s unique singing or rapping voice.”

Daily newsletters straight to your inbox

Sign Up

More From Pro