I keep hearing that AI and robots are taking away all of the jobs. Up until recently I wasn’t too concerned.  After all, I’m not interesting in working at Wendy’s or in an Amazon warehouse.  I’m a CREATIVE person, dammit!  I’m an artist who makes a living off of my human intuition!

At least, that used to be my thinking.

While it is true that art is a special case when it comes to AI, there is an intersection where all bets are off: the corner of ‘Art’ and ‘Commerce’.  But let me back up a little…

When I began my career in audio post, sound effects and music libraries were on vinyl records.  We recorded VO’s onto 1/4 tape and edited them with a single-edge razor blade. (Cutting coke wasn’t the ONLY reason studios always had them lying around.)

The voice actor was always right there with me and my client.  If they weren’t, we directed the session over the (land line) telephone and received the raw tape the following day in FedEx envelope. Home studios?  I remember hearing that Steve Miller had a cool one in Idaho.  Audio post was ‘manual’ labor;  Many were the nights I would come home with a headache and sore shoulders from hours of Benihana-chef-like manipulations of tape reels and razor blades.

A few years later the company I worked for made the move to it’s first Digital Audio Workstation (DAW).  A competing studio had started making waves with their AMS AudioFile, one of the first generation of digital recording/editing devices.

We responded by getting our hands on a New England Digital PostPro.  It was about the size of a refrigerator, and had a black and white interface, limited tracks and no ability to manipulate audio other than to edit it and move it around.  I was in heaven.  It was TOTALLY worth the $100,000 price tag.  It was THAT amazing!

Ten years later GarageBand was introduced by Apple.  It was free.

Once the digital genie was out of the bottle, it didn’t take it very long to grow very big, very fast.  If you were born in the mid-80’s to 2000, this pace of change feels normal to you, much as the proverbial fish doesn’t realize just how wet it is.  To a contemporary composer, sound designer or video editor, progress feels incremental.  More evolutionary than revolutionary.  But here’s the thing. It is natural for our monkey brains to assume that the rate of change in the near future looks pretty much like the rate of change in recent past.  However, as is usually the case, our monkey brains would be dead wrong.

Consider these two lil factoids:

Factoid 1: According the The Economist, the computing power of the then-new iPhones sold over the weekend of their release in September 2014 represented 25 times the computing power that was available to the entire world in 1995. (Factoid 1.1: My first session on that $100,000 New England Digital PostPro was in February 1989.)

Factoid 2:  According to Erik Brynjolfsson and Andrew McAfee, “Two years from the moment you read these words, the planet will add more computer power than it did in all previous history. By the late 2030s there will likely be a thousandfold increase in computer power from where we are today.” *

Chart_AudioPost
Oh.  THAT’S what they mean by ‘exponential growth’

I can hear you thinking, “So?  What’s that got to do with ME?  I’m a CREATIVE person, dammit!  I’m an artist who makes a living off of my human intuition! “

Here’s what it has to do with you, oh creative one:  This explosion in computational power that is staring us in the face corresponds with a relatively new computer trick – they can now learn and teach themselves.  They can absorb millions of examples of something in the blink of an eye, analyze them, break em down, see what works, what doesn’t, try new things and strive for improvement.  You know, kind of like a human artist, but faster and without the need for sleep, caffeine or a bad breakup for inspiration.

Here are a couple examples for you: 

You can now commission a software algorithm to create a piece of custom music.

In a couple minutes.

For next to nothing.

JukeDeck was last year’s TechCrunch Disrupt Startup Battlefield award winner.  And with good reason.  Countless composers make at least a side income in licensing fees by creating basically soulless, forgettable, yet emotionally appropriate scores for corporate videos, explainers, Kickstarter videos and the like.  Well, guess what is pretty darn good at creating soulless, forgettable music?  That’s right.  AI.  At JukeDeck you enter a few parameters regarding style, instrumentation, emotional color and length and then sit back while AI composes a track for you.  Don’t like it? Try again.  Like it? It’s your’s to use for seven bucks.  Twenty two bucks if you’re a big company like Amazon or Gawker.  Want to own it outright?  I mean, ‘own it’ as in use it forever however you like, or even license it to others (who presumably lack internet access)?  That will set you back about $200.

Here’s another one.  Mastering.  Let’s say you’re in a righteous band, you’ve done the whole BandCamp thing and raised enough on Indigogo to record your first EP.  You know enough to not skip the all-important mastering process, but that costs money, and it’s hard to explain what mastering is to the people you’re hoping to squeeze money from.  Fortunately, your roommate went to school with someone who knows a guy who does mastering for local bands, so you hand him your hard earned (or begged for) money and hope for the best…. OR…you head on over to LandR.com.  There, self-learning AI algorithms can take your music file, analyze it, and then apply corrective EQ, multi-band compression, stereo enhancement, etc.  based on the musical genre, the unique characteristics of your track and the desired playback mediums.  They have a monthly or annual ‘all you can eat’ subscription model, but it basically comes down to a few bucks per track.

Once again I can hear you thinking, “Whatever.  Robot music might be good enough for crappy corporate videos, but it will never replace the human nuance and soulfulness of a real human composer!”  or  “Fine, but there’s nothing like the human collaboration of sitting in a mastering suite with a seasoned professional.”

My rebuttal?  It’s two-fold.

OK, maybe three-fold.

First of all, you said it yourself: ‘good enough’.  There is already a big swath of the music market that will settle for less if the price is next to free.  That market is now effectively eliminated for said human composers and mastering engineers.  Forever.  Don’t believe me? “Voice talent suck as engineers, and will NEVER be able to replace my skills, equipment and fancy floating-floor sound booth with their sad, little home studios!”, said countless debt-burdened studio owners who have since seen a sizable percentage of their billings go away never to return, thanks to Guitar Center and a couple YouTube videos.  There are a few more examples I can think of, but I won’t belabor the point.

Second of all, AI will continue to get better at accurately simulating the human compositional process, and at cross-referencing thousands of examples of well-mastered music.  It will get better quickly, and it will get better continuously.  Look at that right side of that chart of computational power again if you doubt this.

And thirdly of all, even your intern can’t afford to work THAT cheap! The less these services cost, the more is left in the budget for other things.  Or more likely, the smaller the budgets will get.  Tis inevitable.

Then there’s that ‘instant gratification’ thing.

By the way, AI is also making inroads in video editing and sound design.  Can mixing, motion graphics and color-correction be far behind?  (Hint: No. They can’t.)

Now, before you get all up in my comments section with your venom and vitriol, I will go on record here and state that of COURSE AI will not completely replace skilled artists in these and other creative mediums!  It will actually create some awesome helpers.  (Video workstations that can transcribe the audio from 7 hours of interview footage and auto-create relevant clips sorted and searchable by key-phrases? Yes, please!)  But AI WILL replace a lot of the work currently done at the bottom of the food chain – draining important revenue and job opportunities from the pond along the way.

In other words, Danny Elfman will continue to score movies, but the nameless composer of the music used in that video explaining how your benefits package is being ‘modified’? For her, things are about to get a lot more challenging.

*Erik Brynjolfsson and Andrew McAfee, The Second Machine Age (New York: W. W. Norton, 2014), p. 251