So, you’re in the explore section of a music streaming app. You’re hands-free, running on a treadmill with the music forming the movement of your workout. Your AirPods stay firmly in your ear as you work up a sweat. Wait, this new song is pretty good. It has great reverb and the beat is pretty catchy. Who’s the artist? Wait, it’s not human? Oh, okay.
The next song starts playing. It’s Taylor Swift singing a song about war and peace. Wait, Taylor Swift didn’t sing a song about war and peace, you should know, you’re a mega-Taylor Swift fan! Oh well, it’s AI again.
How did this happen? How did AI take over music? Already, there’s a fear that AI would replace people’s jobs.
Remember in the early 2000s when Napster allowed you to illegally download music? That was one of the biggest examples of disruptive tech shaking up the business. And now, here’s another disruption.
You’re going to see a lot of AI-generated songs listed on streaming services and all of it is going to redefine what ownership, musical creativity and attribution could look like. Many of the songs that people have been hearing on Tiktok, where fans take existing songs and use software to make them sound like other artists, like Drake. An AI-generated track with Drake and The Weeknd recently went viral on social media, but was removed from streaming platforms. The University of Würzburg tried this out by developing an AI called the Mozart Jukebox to visualize Mozart-esque music for the 100th Mozart Festival. There’s, even, a tool by Google called MusicLM due to be released, where hit pop songs could be created entirely by AI.
Anyone can list a song they worked on, on a streaming service for free. That way, there’s been a bit of democratization as to who an artist can be and that they can come from anywhere.
Production houses and music libraries that have a mass music-as-a-service business may completely go the AI route. And stock music may also go down the AI path, but that would just mean that the music market would be flooded with boring, monotonous and just-okay music.
But, if you’re a singer and you don’t want to go the songwriting route, can you get the lyrics off Bard or ChatGPT and start working on the tunes from there? So, does that compromise work authenticity and work ethic? Or if you have the lyrics, but turn to AI to create the tune, does this make music production more available to people who don’t have musical composition expertise? Would you be sharing credit with the AI? Does this redefine the language between musical creativity and language?
There’s an argument on either side: AI-generated lyrics could sound emotionless and lack depth, but AI-generated music could allow songwriters to better express themselves and even help them out of writer’s block.
Of course, this does raise the issue of copyright infringement. Some are trying to figure it out. For example, AI music generator Soundraw is exclusively trained on music that’s created in-house by music producers.
What about artists whose music has been used to train AI models? Who gets paid there? Would compensation be enough?
Nonetheless, people do believe AI is the way forward for music. Even renowned DJ David Guetta has chipped in to say that AI is the future of music, using AI sites to create lyrics and raps of Eminem. That being said, he opines that taste is not going to be replaced and there are certain emotions that people may want to hear from artists, but AI is like an instrument that would lead to musical revolutions.
For many folks, writing songs is therapeutic. Maybe, it’ll be a bit more mechanical for AI.
But, all that matters is: will the hills be AI with the sound of music? And will you be moved? Will your new favourite pop star be a generative AI tool? Is it time for music to face the music?
