By Gary Symons
TLL Editor in Chief
Artificial intelligence is turning the licensing world on its head, and this week saw major announcements from YouTube and the artist Grimes.
The Google-owned YouTube streaming service has licensed the voices of John Legend and Sia for AI-created soundtracks, while Grimes’ company Create Safe is creating a new tool for publishing and distributing music.
Both of these initiatives have the potential to radically change music creation, distribution and licensing.
Let’s dive in first to the big move by YouTube and Google, which also owns the generative AI division Deep Mind, which created a music generation bot called Lyria.
YouTube announced it has teamed up with a small group of music artists to offer AI-generated versions of their singing voices as soundtracks for creator videos. It’s doing that by using a music generation model created by Google’s AI division to produce unique 30-second clips in a limited trial run.
The artists included in this trial include Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Sia, T-Pain, Troye Sivan and Papoose.
On the other side of this trial run, the new Dream Track program has been opened to a small group of US-based creators using YouTube’s Shorts feature, developed to match the challenge posed by TikTok and Instagram Reels.
In a blogpost, YouTube said creators would be able to produce a 30-second soundtrack by typing a text prompt.
“By simply typing an idea into the creation prompt and selecting a participating artist that appears in the carousel, an original Shorts soundtrack featuring the AI-generated voice of that artist will be produced for the creator to use in their Short,” wrote YouTube executives Lyor Cohen and Toni Reid.
The prompts don’t have to be that detailed, and seem easy enough for virtually anyone to turn out a song with a prompt like this one, that was shown on YouTube: “A ballad about how opposites attract, upbeat acoustic.”
In the example, that simple command turned out a short clip from an AI version of Puth, singing, “baby we’ve got nothing in common.” You can see the example in the video below.
The stunning thing about this is that the voice, melody, instruments and lyrics are all generated by the AI bot Lyria.
The voice, instrumentation and lyrics in the Dream Track experiment are all AI-generated. YouTube has limited access to the tool to about 100 Shorts creators and creatives in the US, but Shorts users can then deploy and remix the track themselves once it appears by using the “use this sound” and “cut this video” features.
YouTube’s announcement also said the company is hoping to add features that make it easier for creators to get the music results they want, such as the ability to create a song just by humming a tune and converting it into melodies based on the genre you want.
All of this comes against a general uproar in the music industry over the potential for AI to violate the copyright of human musicians. For example, the Canadian hip hop artists Drake and the Weeknd had their voices used in the release of a music video called Heart on my Sleeve, which they and their record company Universal Music Group (UMG) charged was not licensed.
UMG is also suing the AI company Anthropic for creating lyrics that are almost identical to those in songs like Gloria Gaynor’s ‘I Will Survive’, and ‘Roar’ by Katy Perry.
Grimes and Create Safe Launching AI Music Generator Called TRINITI
Perhaps in response to the uproar over creator rights, the Canadian musician Grimes is working with the new company CreateSafe, which just raised $4.6 million to launch a generative AI tool called TRINITI.
CreateSafe describes TRINITI as “an artistic intelligence platform that will enable artists, producers, and rights holders to shape the next generation of creativity.”
However, the point of this new tool appears to be based on giving people not only the ability to create music, but also help those creators release their music through a “Record Deal Simulator.”
“CreateSafe’s TRINITI is a comprehensive GenAI music platform,” the company says. “It powers creation, publishing, administration, distribution, and marketing of music, giving artists and their teams unprecedented power to generate ideas and bring them to the public. TRINITI aims to evolve the music industry from its current algorithmic obsession to generating new, artist-focused business models that will break down archaic, imbalanced, and murky industry standards.”
Grimes is the one-time girlfriend of tech entrepreneur Elon Musk, who is also dabbling rather heavily in the AI space, even though he has previously warned that AI has the potential to end human life on earth. Grimes, a musician herself, has joined CreateSafe’s advisory board to continue to experiment and develop new use cases for collaborating with AI.
She started that work earlier this year, and in May, Grimes made headlines for cloning her voice and enabling other musicians and artists to use her voice in new original songs.
Powered by TRINITI, Grimes enabled artists to distribute these works to all major streaming services. Over 1,000 songs have been created since its launch and led to her being named one of Time Magazine’s “100 Most Influential People in AI 2023.”
It’s also worth noting that CreateSafe was co-founded by Grimes’ manager, Daouda Leonard.
What TRINITI and Dream Track have in common is that they are both aimed at creating legal ways for AI to create and monetize music. While Dream Track essentially licenses the voices of various artists, TRINITI goes much farther, essentially making it easier for artists and rights holders to license and monetize their data in generative AI models, while providing an interface to manage consent, credit and compensation.
Actual musicians might point out that creators using AI are not actually creating music themselves, but CreateSafe argues they are simply creating music in a new way.
“Just as an artist becomes a practiced instrumentalist or skilled vocalist, they can train TRINITI to effectively communicate their musical and artistic vision,” the company says. “TRINITI opens an age of multiplayer collaboration, allowing independent creators to make and release songs with some of their favorite artists and producers.”
Must Read: Universal Music’s Grand Experiments With AI and Music Streaming Payments
Leonard says the company’s next phase begins now, as the studio is giving access to artists, producers, songwriters, managers, and rights holders across the world to generate new musical works. He also says the technology will be used to protect the rights of musical artists.
“Music is the core of humankind,” says Leonard. “However, the story of music as a profession has been corrupted by middle men, who have misguided the industry while taking money from artists. For a few years, we’ve been saying that we are building the operating system for the new music business. With AI, it’s possible to fulfill that promise.
“We want to pioneer the age of exponential creativity and give power back to creators. With TRINITI, you can turn inspiration into a song and set of visuals. That music gets distributed to DSPs, a marketing plan can be generated, and all of the business on the backend can be easily managed. This whole process takes seconds.”
Grimes also says the point is not to harm human musicians, but to better reward them for their work.
“As a team we’d always discussed finding novel ways of wealth redistribution via art,” says Grimes. “We immediately hopped onto blockchain tech because of the new possibilities for distribution, cutting out middle men, etc. Throwing generative music into the picture and removing all our label strings so we can reward derivative music – combined with everything we’d been working towards the last few years with blockchain – allowed a unique approach to distribution.
“There’s a lot to talk about but ultimately, art generates so much money as an industry and artists see so little of it. A lot of people talk about abundance as one of the main end goals of tech, acceleration, AI, (but) for us the first step is actually figuring out how to remove friction from the process of getting resources into artists’ hands.”
Why You Need to Watch the Getty Images Lawsuit Against Stability AI
All of that said, a major question is what will happen to highly skilled musicians, if people with limited or even no skills can generate music with a simple text prompt. That also raises legal questions, as the US Patent and Trademark Office recently won a court case affirming the principle that only human works of art can be copyrighted.
In a case heard in August in a US District court in Washington, D.C., Judge Beryl Howell supported the US Copyright Office’s rejection of an application filed by computer scientist Stephen Thaler on behalf of his DABUS system. The art work ‘A Recent Entrance to Paradise’, was said by Thaler to have been created entirely by his AI system DABUS.
The case is a critical one for the future of human creators, as well as licensing companies, both of whom could see their creations replaced by cheaply generated art, videos and written works from AI bots. As one example, the licensing company Getty Images has sued the generative AI company Stability AI for copyright infringement, charging the company trained its AI bot by copying images from its database.
The case follows a key decision by the US Copyright Office. As reported in TLL in March this year, the Copyright Office issued a new policy that says works created with the help of artificial intelligence may be copyrighted, but only if a human being is the guiding mind in the creation of that work. In other words, AI can be used to assist an artist or writer, but cannot replace them.
That raises the critical question of whether ‘creators’ using text prompts to generate music and entire songs from a generative AI bot will be able to get copyright protection, and therefore, whether they will be able to monetize that work.