A live performance from T-Pain, one of the nine artists who are participating in YouTube’s Dream Track beta. Photo Credit: Daniel Benavides
Days after cracking down on certain artificial intelligence videos, YouTube has officially unveiled Dream Track, an AI “experiment” through which creators can use select artists’ soundalike vocals in Shorts.
The Google-owned video-sharing platform, which only recently brought The Beatles’ catalog as well as top music charts to Shorts, formally announced its latest AI-music feature today. Initially outlined in reports last month, Dream Track is powered by Google DeepMind’s Lyria music-generation model and has gone live for “a small group of select US creators.”
Now, these individuals can (after “typing an idea into the creation prompt”) generate up to 30-second song clips featuring AI depictions of artist voices. At present, the trial has nine participating acts, nearly half of whom are signed to Warner Music: Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Papoose, Sia, T-Pain, and Troye Sivan.
While Shorts resulting from the project will likely begin making waves sooner rather than later, YouTube’s released a brief video showcasing the technology at hand. Additionally, YouTube, also in connection with the “AI Incubator” that it launched back in August, is developing a collection of tools designed to enable artists to “bolster their creative process.”
“We’re developing prospective tools that could bring these possibilities to life and Music AI Incubator participants will be able to test them out later this year,” the platform disclosed today of the relevant AI offerings’ quick-approaching debut.
Needless to say, Dream Track could lay the groundwork for the wider release of AI tools that mimic far more than nine artist voices.
Of course, significant questions remain about the long-term commercial impact of AI music’s growing prevalence. Bearing in mind this point and the relative prominence of popular-but-unauthorized soundalike efforts, YouTube further took the opportunity to reiterate that it’s “developing sensible and sustainable controls, monetization and attribution frameworks.”
Warner Music Group CEO (and former YouTube chief business officer) Robert Kyncl spoke at length about the Dream Track deal during his company’s Q3 2023 earnings call, while Universal Music Group reached out to Digital Music News with comments from its own CEO, Lucian Grainge, and participating artists.
“We have a fundamental responsibility to our artists to ensure the digital ecosystem evolves to protect them and their work against unauthorized exploitation, including by generative AI platforms,” Grainge said.
“At the same time, we must help artists achieve their greatest creative and commercial potential – in part by providing them access to the kind of opportunities and cutting-edge creative tools made possible by AI.
“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. This is not a time for passivity.
“Only with active, constructive and deep engagement, can we build a mutually successful future together,” concluded the Universal Music head, whose company’s developing a “music-centric wellness app” and scored an AI-focused partnership with BandLab Technologies last month.
Lastly, here are the full comments that Kyncl delivered about the YouTube Dream Track AI agreement during Warner Music’s third-quarter earnings call.
“I’d like to actually point out the significance of this, which is – imagine in [the] early 2000s if the file-sharing companies came to the music industry and said, ‘Would you like to experiment with this new tool that we built? And see how it impacts the industry and how we can work together.’
“It would have been incredible. Obviously, that didn’t happen. So this is the first time that a large platform, at a massive scale, that has new tools at its disposal, is proactively reaching out to its partners to test and learn. And I just want to underscore the significance of this kind of engagement, and sort of the orderly fashion in which this is happening.
“And I really applaud YouTube, DeepMind, all of Google, and our counterparts in the industry for participating in this because this is the right way to engage. Whenever I say responsible engagement with our partners, this is precisely what I mean.
“So we’re excited about it, we’re excited to learn from it, and together with them then develop a great blueprint for how things should work, but develop it based on learnings. More broadly, the way I think about our engagement on AI and what we practice is along the following lines.
“We have three constituents. One, which is the generative-AI engines, right? So whether it’s DeepMind, Anthropic, LaMDA, etc. And there, obviously that’s where it begins. And there, our efforts in the music industry are focused on making sure that the [AI developers are] licensing content for training, they’re keeping records of inputs so that provenance can be tracked and then there’s watermarking of the content.
“The second group is the platforms where most of the content, irrespective of where it will be created and by which tools, will end up because people who are creating will want views or streams or, you know, lots of user engagement. And with the platforms – so YouTube, TikTok, Instagram, etc., obviously those are the platforms, Spotify – we’re focused on three things, which is control, attribution, and monetization.
“And all of those wrapped in choice, you know, for artists and songwriters, making sure that they have a choice. And we have a blueprint from all of our work on user-generated content over the past 15 years or so, which created a multi-billion-dollar industry on an annual basis for the music industry. So we just need to now write the fine print for the AI age together with them.
“And then the third sort of constituents is governments. And over there, we are both through our trade organizations, as well as ourselves, working hard to make sure that regulation around AI respects the creative industries, the music industry specifically from our standpoint, that licensing for training is required, and also that name, image, likeness, and voice is afforded the same protection as copyright.
“And I myself, I personally have spent time over the last month with leading politicians on these issues and regulators in London, Brussels, Tokyo, and a few others, D.C. and a few other cities around the world. So lots of effort underway, but I’m really excited and positive about the YouTube beta.”