John Legend is among the nine musicians who have agreed to take part in YouTube’s innovative Dream Tracks project.
YouTube has introduced a new experimental feature that enables users to craft short, original songs utilizing AI-generated voice replicas of well-known artists.
Named “Dream Track,” this AI-powered tool is initially accessible to a select group of creators, as detailed in a recent blog post by YouTube. The project has garnered participation from nine artists, including Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Papoose, Sia, T-Pain, and Troye Sivan.
Dream Track is designed to produce unique songs, each lasting up to 30 seconds, tailored for YouTube Shorts. These Shorts are concise, vertical videos akin to those found on platforms like TikTok and Instagram Reels. Users can input a brief description of the desired sound, such as “an upbeat acoustic ballad about the attraction of opposites,” and select one of the nine featured artists. Subsequently, Dream Track will generate a snippet of a new song that fits the specified criteria, sung in the AI-generated voice of the chosen artist.
In the initial phase, this experiment aims to explore how the technology can facilitate deeper connections between artists, creators, and their fan base, according to the blog post.
The tool leverages Lyria, Google DeepMind’s AI music generation model. Notably, both YouTube and Google DeepMind are subsidiaries of Alphabet, the same parent company. The generated songs will contain an embedded watermark indicating their AI origin, although this watermark is imperceptible to the human ear, as outlined in a Google DeepMind blog post.
In statements provided by YouTube, the artists have expressed their support for the experiment, with some, like Charli XCX, expressing reservations about AI’s potential impact on the music industry.
“A.I. is poised to revolutionize the world and the music industry in ways that are not yet fully comprehended,” Charli XCX remarked. “This experiment will provide a glimpse into the creative possibilities that may arise, and I am eager to witness the outcomes.”
The introduction of Dream Track, featuring T-Pain, marks a significant milestone in the realm of YouTube Shorts.
AI-generated music has stirred controversy in recent times. Earlier this year, a creator known as Ghostwriter produced a song titled “Heart on My Sleeve,” featuring AI-generated vocals resembling those of Drake and the Weeknd. Following the song’s viral success, Universal Music Group, the artists’ record label, successfully persuaded YouTube and other music streaming platforms to remove the track due to copyright concerns.
YouTube recently unveiled fresh guidelines governing AI-generated content on the platform. These regulations mandate that creators clearly label their realistic AI-generated videos as such, and YouTube will establish a procedure for the removal of deepfake videos upon request.
In the evolving landscape of AI-generated music, stakeholders such as artists, legal experts, and music platforms are grappling with the ethical and legal ramifications of this technology. Questions persist regarding the implications of copyright laws, remuneration for AI-generated music, and the potential fallout if AI generates contentious lyrics that harm an artist’s reputation.
American musician and DJ BT highlighted the potential risks, stating to NPR’s Chloe Veltman in April, “When it comes to creating vocals, it could be utilized to convey messages diametrically opposed to an individual’s belief system.”