Tackling Fake Drake
Industry Overview and the Implications of Generative Music
A couple of months ago, a revolution started that snuck by our ears and widened our eyes. Sitting in my apartment in Blacksburg, Virginia one evening, my roommate came over to me and turned on a song. My ears instantly recognized the voices of Drake and The Weeknd, two of my favorite artists, but my mind just couldn’t wrap itself around the name of the tune. After this, what initially began as confusion for never having heard this song before (I’m known to be somewhat of a Drake mega-fan) quickly turned into awe as my friend explained to me that this was not actually a product of the two mega-stars.
“Heart on My Sleeve,” as the song was called, had been created entirely by generative AI to produce a mimicry of a track between two of the world’s biggest names in music. “Heart on My Sleeve” represents the most infamous example of a genre of music that has exploded recently: tracks created by amateur producers using generative music platforms.
Although the track was removed at the request of industry titan Universal Music, its mark had already been left. Instantly, comparisons of the generative music technology to other major disruptions in music history were made. Parallels were drawn to inventions like the synthesizer, which revolutionized pop music and accelerated the development of completely new sounds. People were even mentioning the disruption of generative music in the same conversation as streaming itself, which boomed our consumption of music and upended traditional ways of distribution.
Coincidentally, my team and I at The Consulting Group At Virginia Tech had been wrapping up a project with Eola Arts Incubator to develop a business analysis and industry overview regarding AI. Specifically, they wanted to know how generative music could help their business and aid their artists. Now more than ever, artists face an exciting – albeit daunting – entrance into a world that will soon to be dominated by AI. Generative music opens up a world of possibilities for all ranges of musicians. In order to seize these opportunities, however, it is critical that artists embrace this new era with enthusiasm, creativity, and a whole lot of humility.
The Dawn of Generative Music
OpenAI, the company that brought the world ChatGPT, is currently one of the largest players in the generative music space with their Jukebox platform. Jukebox is the purest example of a generative music platform. Simply upload an audio file, pick a genre and artist, and Jukebox will generate a piece of music unique to your inputs. Easy right? Although the steps to creating music through Jukebox are simple, the complexity of how it processes and outputs music makes it unrealistic for most artists to use. Jukebox models music directly as raw audio by mirroring and learning from a song's timesteps. A standard, four minute song has over ten million of these timesteps, resulting in one minute of audio taking nine hours for the platform to generate.
Platforms like Aiva, which is trained through a database of millions of audio files, are not nearly as time-prohibitive and stand as a great alternative to Jukebox. With Aiva, a user inputs their own audio files from a song they wish to model along with a preferred genre, artist, tempo, and other parameters. Aiva then creates a new song that – as opposed to Jukebox – creators can use however they see fit. Paying Ava's monthly fee gives creators full monetization rights over the music they create, and users are able to upload them to any and all platforms that they wish.
What if artists were not necessary in the first place? Companies like Aimi, a streaming platform that allows users to influence the music they are listening to through inputs, flips the entire concept of music on its head. Aimi has various pre-set streams in different genres that play continuous loops of music. Users can rate the song they are listening to as a means of tailoring the sound exactly to their liking. Operating similar to Spotify’s “freemium” model, users have access to thirty minutes of free streaming a day before being prompted to pay $4.99/month to continue with unlimited music.
Whether or not platforms like Aimi catch on with the public, the concept of allowing total control of the listening experience is far ahead of the industry standard. In a world where personalization is becoming the norm rather than the exception, traditional streaming platforms such as Spotify and Apple Music will be forced to innovate using technology like Aimi’s in order to serve their listeners.
The Democratization of Creativity
Up until generative music platforms became available to the public, the ability to create good, quality music was entirely up to the musical ability one had. Now, individuals must simply be armed with a musical idea and the ability to train a generative music platform which, as we have seen with ChatGPT, is easy for one to learn to do. This, in my opinion, is the biggest reason why generative music is so exciting: it has the potential to allow pure creativity to usurp traditional music talent as the new key driver for success. As generative music technology continues to improve, the music playing field will continue to level. While some may see this as an attack on artists creating music traditionally, those who embrace the rise of generative music will find that creativity is the true winner, bar none.
As a scenario: say you are an aspiring musician with great vocal talent and no other traditional musical skills. One day, a melody for a song comes to you in your mind, and you do not know what to do with it as you do not know how to play any instruments or produce music. Prior to the advent of generative music, you would have had to find someone who does have these skills or take up learning them, which for many people can be almost impossible after a certain age. Now with technologies like Aiva or Jukebox, you can upload a file of your voice singing the melody and find yourself at the receiving end of a beautiful, well-composed song. This is not an attack on traditional artists but rather a celebration of creativity and the love of music. While those who want to delve into the minutia of music production can still do so, these platforms can get you 99% of the way there without the barriers of traditional music knowledge. In my view, generative music is a sure-fire net positive for the industry, as it provides the missing link that can enable you, me, or anyone else with a spark of creativity to bring their musical vision to life.
AI Business Models for Musicians
After the release of “Heart on My Sleeve,” many artists began to think about how exactly to grapple with its monumental implications. While stars like Drake and The Weeknd can count on their labels to take down music created by AI that is impersonating them, other independent artists have started to think of ways to capitalize on the trend. Notably, Canadian musician Grimes has offered AI artists a 50/50 split when using her voice in songs. Not only this, but Grimes even released her own software to aid people in creating music in her likeness (must be nice having Elon as your ex, ha). This makes a lot of sense from a business perspective: why not get paid to have people do your work for you? Although major record labels are aggressively pushing back on AI songs created in their artists' voices, deals like the one Grimes is making bear the question: how sustainable is this? When will it get to the point where their own artists begin to question why they do not have complete sovereignty over the very thing that keeps the record label and the music industry afloat: their voices and likeness.
Admittedly, it is unlikely that artists in the category of Drake and The Weeknd would be willing to take up a deal like Grimes’. However, a strong business case can be made for an artist in the twilight years of their career to arrange this sort of deal as a means of continuing to spread their presence. As another scenario, let's take an artist like the rap legend, Ice Cube. Ice Cube is now in the era of his career where he is focused on areas outside of music, such as his successful Hollywood career and various endorsements. While having diversified himself beautifully, he still has a huge fan base solely focused on his music, as is apparent with his 18 million monthly Spotify streamers. An astute businessman, it is fair to assume that Ice Cube would be happy to have music released in his voice without having to spend all the time to write lyrics, record in the studio, and produce. Heck, he could even have control over which songs get released. Busy filming a movie? Take three minutes to listen to this song we made for you and let us know if you want us to release it. Likely, his and other artist’s labels will have people on board whose sole purpose is to cultivate AI generated music for their artists. Artists making money without having to work, labels making money, fans happy. Three-way win.
The Future is Now
While a change in the way music is created and enjoyed can be daunting for creators and listeners alike, the prospect of the adoption of generative AI in the music industry can be an incredibly powerful and positive force when leaned into with an open mind and open ears. Admittedly, there is the concern – as is with many use cases of AI – that smaller players in the field could be displaced. With music, I do not believe that artists themselves should worry but rather those who are involved in the production and mixing of the music. If AI can hear an artist singing their lyrics and then quickly produce a beat that is perfectly tailor-made for the song, one may start to question the need for these middlemen.
While much of what this article has been expounding upon is theoretical in a sense, the one thing that I can say with 100% certainty is this: AI is here to stay. It will not wait for you to catch up to its capabilities. Furthermore, the technology and platforms we have for generative music and generative AI in general will not stop improving exponentially. If you are an artist, embrace this new landscape and dive into it with the mindset of, “how can AI make my artistic process better,” and not with, “how am I going to compete with those who are using it.” To end in the most cliche way possible: the future of music is here. The opportunities that AI can lend to artists are endless, and this should be music to everyone’s ears.
About Marco Ferrara
Marco is a student at Virginia Tech studying Business Information Technology through the Pamplin College of Business. Marco is a member of the Consulting Group at Virginia Tech where he works with local, regional, and multinational clients to develop strategic solutions to complex business processes. He is currently working for PWC in their Consulting Solutions service.