At the rate AI-generated music is progressing, it could be old-fashioned to manually compose songs in as little as a decade. The latest step forward comes with the release of "Break Free," the first single from the album IAMAI, for which all the music is entirely written and produced by artificial intelligence.
Machines have been slowly infiltrating creative fields for several years, but more recently we have seen the pace of what can be achieved using artificial intelligence rapidly accelerating. Last year, Sony's Flow Machine software generated its first pop song, designed in the style of The Beatles. The lyrics of the song were penned by a human composer, who also helped arrange the AI-generated segments of music.
Now YouTube star Taryn Southern has joined forces with Amper, a company that has developed an AI system that can produce professional sounding music in seconds.
For the first single off the upcoming human/AI collaborative album, Taryn used the Amper AI to entirely compose the music from scratch. The only manual inputs Amper needs are beats per minute, rhythm, mood, and style of music. Taryn then penned her own vocal melodies and lyrics, which she recorded and added to Amper's music to make the final product.
"I get to spend more time doing more of what I love – 'conducting' the AI to create a great song, writing lyrics and vocal melodies, and making music videos," explains Taryn regarding her production process.
The final song is frighteningly similar to plenty of pop music out there, so depending on your tastes it is either absolutely awful or touchingly evocative. More definitive is the eerie way the AI has perfectly captured the progression, flow and tone of a modern pop song.
Amper initially developed its AI music composition system as a way for businesses and small content creators to add background music to their projects without delving into the messy world of royalties and licensing. The technology is clearly progressing at a rate where it can now effectively replace human composers across a multitude of tasks.
There is currently a rapidly growing sector of startups working on developing AI-generated music systems. Most are content to sit in the "muzak" world, generating music designed to play in the background of department stores or elevators, but it is clear that this technology can be used by amateur musicians as well.
Just as other technology advances have enabled bedroom musicians to create complex electronic music compositions, this new world of AI-generated music will now allow artists to create entire original songs nearly instantly.
Can't write or play music? No sweat, let the AI make the song. Can't sing in key? That's cool, we have AutoTune for that.
The singer picked the beats per minute, rhythm, mood, and style, then wrote the vocal melodies and lyrics...
Seems to me, that with all of the above done for it, a computer about as sophisticated as an abacus could have done what little there is left to construct a passable song..
When completed, the song turned out to be the very definition of derrivative.......we've got loads of music like that already, thanks.
It would allso have helped if they spent more than a couple of quid on the video...
I don't think pop songwriters need to loose sleep just yet.
I'd have to agree with some others commenting on several points they brought up.
AI "created" is more like gambling at this point. Or picking lotto numbers.
As a "composer" I believe a lot of humans base the new songs on sounds they heard prior. The human sense of "inspired by" isn't being translated into the AI composition. The AI is simply compiling based on a given selection of scores.
That doesn't mean the AI compiled random score won't sound good or have a great section while another section sucks. But, it can't tell if that section is great, or sucks...
FL Studio for example has a "roll the dice" section for compositions. It will randomly construct a song. Some sound good, others well.. they suck. So you simply click, roll the dice again until you find something you like. And it too is based on given data. Cords, tempo, style, etc etc.
Point is, until AI can show it "likes it" or the music it created makes it "feel" something... It's just compiling random notes and stringing then together based on a given equation of requested results. That's not intellect, that's math.
But as a writer, I would rather compete against only humans.