AI, Streaming, Royalties, Sustainability… Blogmas #9

The powers that be in British broadcasting say that they’ll be going to streaming within the decade. The broadcasts will be switched off. This is a bit (a lot) of a worry for a great deal of media composers who write for telly programmes because ROYALTIES.

Composers get royalties every time their music gets broadcast in connection (‘synchronised’) with a tv show. Composers will also usually (sometimes) (occasionally) receive royalties when a show of theirs gets streamed BUT the amounts are paltry in comparison with broadcast rates.

So if the broadcasts get turned into streams… TV composers will get less money.

The only option is to charge more upfront. So a composer has to guess how much a piece of music might earn in royalties over its lifetime (which can run into the hundreds of thousands) and then ask a commissioner for that.

Hahahahah yeah that’s not going to happen.

The problem with composers not getting royalties is that the drip drip of extra cash during periods of time between projects can become a bit of a lifesaver. It’s not simply composers being ‘money-grabbing’. It’s one of those essential streams that makes the career of composition sustainable for little nobody composers like me.

What may happen is that only composers who can afford to work for little or nothing will continue to take the work, and composers like me from working-class backgrounds without the bank of mum and dad to stump them up the rent, heat, light and food will be forced out of the industry early in their careers.

It’s unsustainable and elitist.

Sidebar: Games composers may look on in perplexity – royalties from game music are so vanishingly small as to be non-existent. But game composers absolutely do get paid more up-front in commissions, on average.

Here’s another thing – stick with me here, it gets a bit… fiddly…

There is a recent(-ish) fashion in broadcast tv that the commissioner will request the publishing rights. This equates to 50% of the total generated performance royalties. The other 50% is known as the writer’s share and will always get paid to the composer (in UK law anyway, don’t talk to me about the Wild West of US law, I can’t even).

But in the good old days, composers would assign their publishing share to a publisher of their own choosing. These publishers (if they’re one of the good ones) will already have a working relationship with the composer and will likely be more of a partner in the composer’s career.

(Alternatively, the composer should also have the option to keep the publisher’s share for themselves, if they’re up for doing some additional admin from time to time, chasing up unpaid performance royalties, and finding ways to ‘exploit’ the music in the composer’s catalogue for further uses and further income.)

If assigned to a third-party publisher, the deals would be anywhere from 10-30% of the total amount of royalties, so composers would be paid the remainder of the publisher’s share by the publisher.

These days, commissioners often don’t give you the option. It’s ‘we keep the publishing or you don’t get the gig’. And of course, more often than not, we composers really do want the gig. The job is ace, but our enthusiasm becomes a means of exploitation.

Chipping away at the edges, composers attempt to get by with smaller and smaller amounts until they’re forced out of the industry because the pay is, you guessed it, unsustainable…

Still with me? Huzzah!

And yet another thing:

Unless you’ve been under a rock (if you have, lucky you. Sounds peaceful, can I join?) you may have noticed the sudden explosion in interest in the latest AI tools. GPT Chat can put together decent tracts of realistically human-sounding text (even if it is a little bland). Stable Diffusion and DALL-E make uncanny artworks.

There are attempts at music AI but they are barely passable at the mo and can’t compete with a human. That will change. I give it less than 10 years.

Not least that there’s the possibility of work lost by creators to AI; however, the more pressing problems with AI are manifold. Most significant are ethical issues with the datasets that the AI learns from.

Issue 1: These datasets are huge and the information, be it text, images or music must come from somewhere, created by many, many humans over the course of all of recorded history.

Will creators who are alive now, or whose copyright has yet to expire, be credited and paid royalties? Hah! I’ll believe it when I see it.

Will creators be able to stop their work from being scraped from the web and included in these datasets? Too late for that, they’ve already done it! Again, it’s vanishingly unlikely. The only way to be sure is to not post anything online, a painfully unrealistic option for a jobbing creative.

Issue 2: this one’s more insidious, and I’ve yet to see it discussed much anywhere. Like the dirty little not-so-secret of fast-fashion, the sweatshop, AI datasets have to be tagged manually and are done so by armies of underpaid, undervalued, oppressed humans. If these people were paid properly, the costs of using the datasets would likely be prohibitive, I imagine.

So here we are in 2022: The genie’s out the bottle with AI. We’re moving to post-broadcast, streaming-only media consumption. Whether you look at the dwindling royalty system for creators, rights grabs by commissioners, the many dubious ethics of creating datasets, or any other de-valuing of a human’s creative output in the 21st century, one thing is true: I’ll say it again, it is all unsustainable.

Oh, human beings. I mean, come on. We have such potential. And yet, again and again, en masse, we really do let the side down. Le sigh.


I promise the next post (at least) will attempt to be more optimistic!


Image: I made this with DALL-E before I knew the ethical quandaries of AI Art generation. But it kind of sums up the weary desolation I feel sometimes when I get out of the studio and look out onto the lay of the music business land every now and again. Before I scuttle back…

Discover more from Heather Fenoughty

Subscribe now to keep reading and get access to the full archive.

Continue reading