Heralded as the ChatGPT for music, Suno AI is the latest iteration of generative artificial intelligence to flood social feeds, wowing users with its (ahem) lyrical prowess.
Plug in the musical style you want, a genre and a prompt for lyrics and Suno can spit out a full song for you in a matter of seconds.
The business has been around for two years, formulated by a group of machine learning experts in Cambridge who struck an interest in audio, according to a profile in Rolling Stone last month.
From the outset, making silly songs is slightly addictive. The lyrics might seem shallow and soulless, but they’re also often hilarious.
Asked for a power ballad about a morning coffee, Suno came up with:
Coffee, you’re my fuel for the soul (oh-oh)
Without you, I’m feeling so cold (feeling so cold)
Coffee, you ignite my fire within
With each sip, a brand new day begins (oh-oh-oh)
When asked for a dance pop song about unrequited love, it spat out:
I’m watching you from the shadows of my mind
But you’ll never know the feelings I hide
In these neon dreams I’m lost in the night
But my heart keeps hoping – maybe one day you’ll be mine
And in response to a request for a jingle about Guardian Australia, it went with:
From the shores of Sydney to Darwin’s embrace
The stories we tell leave no room for disgrace
Investigative journalism, always on the go
Guardian Australia, letting truth flow
That’s what you get if you choose to have Suno generate the (often quite stilted) lyrics. If you want to put in slightly more effort, you can enter your own lyrics along with a genre and see what happens.
One prompt engineer last week tweeted a sad girl song using the text of the MIT software licence as the lyrics and it’s surprisingly catchy.
AI-generated sad girl with piano performs the text of the MIT License pic.twitter.com/h5wdMuNUdm
— Riley Goodside (@goodside) April 4, 2024
As you can see the quality of the song will vary, but you can refine it and create up to 10 songs a day on the free account (which can’t be used for commercial reasons). For $10 a month you can make up to 500 songs and also commercialise them by uploading them to services such as Spotify or Apple Music.
One of Suno’s co-founders, Mikey Shulman, told Rolling Stone that the aim of Suno is not to replace artists, but to make the app fun and democratise the creation of music by making it more accessible to others.
But it doesn’t take much to hear similarities to songs you know. While you can’t type in a specific artist’s name to create a song in their style, you can definitely steer towards it via your prompts.
The tool can reportedly identify when lyrics you want it to use are subject to copyright. The company’s terms of service say that permission must be sought if you plan to use copyrighted lyrics.
But the main point of contention with the tool is whether or not it was trained on copyrighted material. Guardian Australia asked Suno to clarify this but did not get a reply by deadline.
Ed Newton-Rex, a composer and the CEO of Fairly Trained, an organisation founded to license content provided with consent to companies that train AI, posted on X last week several examples where, without naming certain artists, songs resembling those by Eminem, Queen and others were generated by Suno AI.
Artists are already concerned about what the effect of AI-generated music will mean for their industry. Elvis Costello, REM, Billie Eilish, Katy Perry and Jon Bon Jovi signed an open letter last week calling for AI companies to pledge not to develop technology that undermines or replaces the roles of songwriters or artists.
According to the letter, they fear that AI trained on their work could be used to create massive quantities of AI-generated sound and images that could replace human work and dilute royalty pools.
One of Suno AI’s investors, Antonio Rodriguez, also told Rolling Stone that he invested in the company with the knowledge that record labels could sue. The company said it was in communication with the major labels and sought to respect the work of artists.
A dystopian outcome of a product like Suno would be music streaming sites becoming clogged up with AI songs that sound similar in style to artists already struggling to earn a living from those very same platforms.
Suno AI seems to be aware of this potential outcome, saying an inaudible watermark is embedded into each song it generates so AI music can be identified.
The next question is whether any of the music streamers will put a handbrake on AI-generated music. Spotify, at least so far, seems to be leaning towards allowing AI-generated music that does not directly rip-off artists.
