In Part One of MONTAG’s poke around in the future of music, Joe Sparrow discovered that we’ve been stepping back from the music making progress for a while, and how we’re already consuming music that’s created “robotically.”
In Part Two, having comforted himself that humans are essential to making music, Joe discovers how actually music will be churned out by faceless, automated music factories - and it’ll make the world more musical, more enjoyable and more exciting.
What is life?
Music unites us as people or strengthens us on a human level. Maybe your nights are spent prodding flashing buttons on a Maschine, producing the next banger to soundtrack David Guetta’s next public meltdown; or maybe you just slip on chunky headphones, set Spotify to Private Mode, and lose yourself in the Spice Girls.
So if music is so essentially human, won’t music that’s whelped by a server – not a songwriter – just be rubbish?
Well, no. Everything you know about music – who makes it, how it’s made, how you’ll feel – is already changing, although you might not even notice…
The last time I poked the AI box was to see if technology was ready to take over my creative role and write my MONTAG article for me. It wasn’t an encouraging exploration, and ended with me having to write the damn thing myself after all. Pfft. So much for robots replacing our work, eh?
Except: that piece did end with these words:
“make no mistake, the machines are coming for your job.”
To recap: if you’re a truck driver, now is the time to either start thinking about other jobs, or how you and the other thousands of drivers are going to complain to the dithering politicians afterwards. (And if you’re a smug creative – a phrase which is certainly oxymoronic – note that your job will be gobbled up by technology too.)
But let’s not linger on the impending job-hunt anxiety. What if the future is not all doom and gloom? Automation theory states that when all the simple jobs will be done by robots or AI, us humans have time to get on with doing unique, human things, like creating stuff.
So what if the things that new technology brings us are new types of beauty, unthinkable creative curveballs, and tools that empower us to be super-artists? Here’s how the most universal art form of all - music - will be tweaked, re-shaped and blossom.
Pretend we're dead
Spotify isn't a music company, really: it’s a data company which shuffles around ones and zeros that happen to be mp3 files.
Open Spotify - or any of the identical music streaming apps – and you’re faced with playlists compiled for specific feelings and situations: music to help you focus, songs to work out to, and music to fall asleep too.
While it’s tempting to suggest that clicking on the latter reveals the entirety of Coldplay’s back catalogue, it actually throws up playlists like Peaceful Piano, which are packed with beautiful pieces by critically-acclaimed artists like Nils Frahm, Max Richter, and, erm, Sigimund and Lo Mimeaux.
The latter two have been disdainfully - and, in Trump’s Fake-News world, fashionably - called “Fake Artists”.
These artists are, if you like, human bots: pseudonyms of a few music production teams who grind out music, and lots of it. In bulk, to order, and at a high quality, which is then sold outright to Spotify.
The benefits of this in-bulk music are clear: padding out playlists with these songs in between the expensive “real” artists - who require a royalty fee for every stream - Spotify saves money.
And listeners probably don’t even realise, or care: if you’re using the songs to fall asleep, what’s the difference if Lo Mimeaux isn’t a critically-acclaimed “real” artist? Tinkle those ivories and zen me to sleep, baby.
What Does Your Soul Look Like?
But what does this have to do with future technology and music-making?
It’s this: cheap human-music is the first step to cheaper automated-music, and there are parallels in other businesses.
Consider how you may have made Uber an essential part of your life, and Uber essentially subsidises your fare every time you take a trip. That’s because, one theory states, Uber are waiting until the most expensive part of the process - the human driver - can be replaced, and then the real money rolls in.
Meanwhile, users don’t notice or care: they focus furiously on not drunkenly vomiting in the taxi as they’re driven home from the bar either way.
A cynic would say this mentality, minus the possibility of vomit (unless you’re listening to a Celine Dion album), would apply to music too.
As important as image is, we rarely care about the source of music as long as we enjoy it: very few people in the West listened to K-Pop before Gangham Style exploded in popularity via sheer crazy persuasiveness.
Enjoy Nils Frahm while you can tell the difference.
Computers are only starting to write music, but soon they’ll be writing big hits, and you’ll be singing along. So what is the likelihood that AI can nudge the puny humanoids out of the creative process entirely?
Let’s answer that question with another question: do you loathe jazz or do you merely hate it?
That’s mean, of course: plenty of people love jazz. It’s just that… well, jazz can be shorthand for the kind of tuneless, solipsistic music-meandering that AI-driven music can hopefully banish for good.
The late Tony Wilson – who, as founder of Factory Records discovered Joy Division, New Order, Happy Mondays, and (allegedly) turned down The Smiths because he didn’t like Morrissey - said that “Jazz is the last refuge of the untalented. Jazz musicians enjoy themselves more than anyone listening to them does.”
Let’s assume for a moment that this is true. So where does that leave DeepJazz?
DeepJazz is a jazz machine. Its AI was knocked up in a weekend as part of a hackathon, and it immediately started spitting out the kind of piano-tinklin’, note-meanderin’, taste-dodgin’ jazz made by humans the world over.
Jazz afficionados will rightly turn their nose up at DeepJazz - but to the rest of us it is recognisably jazz.
So if AI is able to make bad jazz, is it not essentially just… making jazz? What is good jazz anyway?
Don’t Believe The Hype?
In 2012, a team at Sony CSL’s research lab in Paris trained an AI programme to listen to pop music, and then asked it to make its own with that it had learned. It produced Daddy’s Car, a song with a structure and lyrics in the style of the Beatles, which was then sung by a poor inconsequential human being.
It’s… fine. It’s not Penny Lane, but it’s not Dig A Pony either.
Recently, Spotify hired the man who is the brains behind it all. Remember their “fake artists” making “fake music?” You can see where this is going.
If a streaming service or label could create a bot that writes mega-hits, we'd all win. Listeners could hear more music that they love, and streaming services would be happy too.
How? If a song as utterly MOR as the Zayn Malik/Taylor Swift snooze-fest I Don't Wanna Live Forever can get over half a billion streams on Spotify, imagine how valuable it would be if Spotify-Bot could write a song that big without having to pay someone for it.
OK. So where does this leave music now? What does it mean for listeners? And what about poor musicians, who get, on average, 0.7¢ per Spotify stream of their song?
Music, despite endless “it’s not like the good old days” hand-wringing, is going to get better. It always does.
If you crossed the ethos of Brian Eno’s Music For Airports and the battery-production ethos of Spotify’s “Fake Artists” you’d not land a million miles away from Muzak™ - the name of the vapid background music you hear in lifts, shopping malls and waiting rooms.
But vapidity is not where music is headed: we don't await a world of panpipes and music that sounds like Enya B-sides. Instead, there’s going to be a lot more music, wherever and whenever you want it - and it’s going to fit the situation you’re in. And that’s cool.
AI is going to make more good songs: whether completely or partially autonomously. Few people panicked that keyboards replaced real orchestras on pop songs.
So why worry too much about the songwriters? Truly good human songwriters will always be treasured for as long as they write amazing music.
Now, if you really want to worry about the future of music, consider a completely human-free process: songs written, produced, and performed autonomously from start to finish.
Something to think about when you’re on the way to a Hatsune Miku gig…