Over the past year, we’ve heard a number of music companies boast about the sheer number of tracks they’ve been able to create using generative AI.
For instance, AI music app Boomy has said its app has created 14.4 million tracks so far.
Statistics like that aren’t music to the ears of many music industry insiders, some of whom have raised concerns about the possibility that AI music will drown out the voices of human creators.
They have some reason to worry: As of last count, there are 120,000 new audio files being uploaded to music streaming services every day.
Now AI music creation platform Mubert is adding its own jaw-dropping number to this data. The company announced Wednesday (July 12) that its AI has generated 100 million tracks – roughly equal to the entire catalog available on Spotify.
Mubert says the audio files were generated “exclusively using licensed music for input.” The company says it has established relationships with music creators who contributed audio samples to its AI, allowing Mubert to create a database of 2.5 million “proprietary sounds” on which it trains its algorithm.
Mubert’s core music-generating product, Mubert Render, has 100,000 monthly active users, the company says. It also offers Mubert Play, a subscription-based streaming service, and Mubert API, a music generator for B2B customers.
Users have generated 56 million of the 100 million tracks that Mubert now has in its library, the company says.
The most common genres of music generated were lo-fi, ambient and chill – something Mubert says makes sense, given that much of the music created on its platform is meant to be an accompaniment to online streams and shows, interviews, shorts and podcasts.
“We are thrilled that Mubert is able to meet the demand for legal and high-quality music to supply the needs of the creator economy,” Alex Mubert, co-founder and co-CEO, said in a statement.
“it is impossible to imagine streams, podcasts, and shows without music, and Mubert allows for the generation of an unlimited amount of music of any duration and any genre, tailored to the needs of the creator economy.”
However, background music for online shows isn’t the only activity of Mubert’s generative AI. The company entered into an agreement with Middle East and North Africa-focused music streaming service Anghami, through which Anghami has created a library of 200,000 songs so far.
Among Anghami’s uses for the Mubert algorithm was a feature called “football cheers,” which was available in Persian Gulf countries during the 2022 World Cup. It allowed users to declare which country they were “cheering for,” and Anghami’s tech would generate a unique song for them, informed by the customer’s user data. Those songs are now hosted on the Anghami server.
Other music services have also reported creating large amounts of music using AI. Tencent Music Entertainment, which operates streaming services in China that boast some 800 million active users, said last year it had generated some 1,000 tracks through AI, one of which had already surpassed 100 million streams.
While many in the industry worry about what such a flood of content could mean for the economic value of music going forward, Mubert’s co-founder, Alexey Kochetkov, argues that generative AI can help address the “winner-takes-all” nature of the music business, “where industry moguls make millions, while new and prospective artists struggle.”
In a guest column for MBW in 2019, Kochetkov wrote that AI “can lend a helping hand in forming a new music industry culture – both qualitatively and quantitatively…
“As long as there is no compromise for AI, the music industry can become a transparent environment where all stakeholders receive equal chances of self-expression and fair conditions for monetization.”
However, some question whether Mubert creates such “fair conditions for monetization.”
In a recent guest column for MBW, Ran Geffen Levy, CEO of Amusica Song Management in Israel, said that AI companies could stand to be the major financial beneficiaries of AI-generated music.
Of Mubert, he wrote: “The licensing agreement provided by Mubert states that Mubert is the sole owner of all economic rights to the remix, such as the so-called ‘master rights’ to the recording, the so-called neighboring/performer rights that may accrue to those who perform on the recording, and the rights to the musical composition that is embodied on the recording.”
Levy asked: “If you take the humans out of the performers’ equation, where does it leave the rest of them? What is the compensation for the humans that write and perform the music? Will musicians receive recognition for their work, or will they remain anonymous and work under NDAs and buyouts? Are we on the verge of creating musician sweatshops? Hey, at some point, AI will be able to replace them all.”
