Will musicians soon join actors on strike against AI?


In 1942, radios went silent. At the height of the war, when the world needed the balm of music most of all, strike action was taken in the US. James C. Petrillo, the union head of the American Federation of Musicians, called upon every ditty-drummer, serenade-singer, and flute-tooter in the land to down tools. When calls for songs to resume and offer a salve to citizens in trying times countered his strike, he insisted: “I am satisfied that if the public of America knew the plight of the musician – knew what he is up against – that the public sentiment would immediately change.”

The main plight that they were up against was the rise of technology. In the 1940s, roughly 60% of the American population went to the cinema once a week. In previous years, each cinema employed a ‘player piano’, as Kurt Vonnegut calls them – ironically, in his novel that satirically muses about the technologically driven obsolescence of humanity. However, when recorded soundtracks could be synced to the talkies, cinemas cleared out their orchestra pits, and thousands of musicians were out of a job. This was indicative of the changing times.

It was also indicative of the double-edged sword of music’s relationship with technology. Without microphones, we’d all still be listening to opera; they gave crooners a chance to literally have their voices heard, embellishing music with a sudden wealth of nuance… and then jukeboxes kicked those performers to the curb, and so on, as the myriad unfurling reverberations of tech took hold of the arts. This led to Petrillo putting his foot down. When he became president of the AFM, he quickly decreed that union musicians would hereby refuse to enter recording studios. This now seems obscene, but at the time, thousands of musicians faced job losses, so he tried to enforce their rights and hit the industry where it hurt.

Nevertheless, Petrillo recognised that stopping recorded music was like fishermen striking against the tide, so he eventually demanded that if things were to go ahead with technology at the fore, then musicians should be paid royalties every time that a record they played on was spun on the radio. It is thanks, in part, to this notion that musicians now get any secondary royalties at all.

This is not the only time that industrial action had called a halt to music when technology began to encroach. As is so often the case, bloody Barry Manilow finds himself as the antagonist in yet another cultural war. In 1982, he was about to head out on tour. In an efficiency bid, he binned off his orchestra and employed a couple of synth players instead. The Musicians Union hated this. So, on the birthday of that dastardly synth pioneer Robert Moog, they decided to push towards outlawing this new musical contraption.

Obviously, they failed in this motion, and the 1980s became the most synth-heavy decade in history. While synth players were banned from the Union until 1997, this did little to curtail the rise. In fact, in a sort of anti-Luddite futurist storm, the facsimile of synthesised sounds became an artistic tenet of the era—musicians used the technology not to mimic a handful of violins with just a few keys but rather to enact some ‘Computer Love’, so to speak.

Alas, that is all ancient history now. We are on the precipice of another great leap in musical technology, and this one oddly seems even more seismic than what has gone before, because rather than reckon with our own governance of how technology should serve us, the question with AI is whether technology will serve us at all or simply enslave us. Part of the reason that sounds so profound and perturbing, aside from the sci-fi overture, is because AI is currently mystified by the fact very few of us know what the hell it really is.

So, I spoke to the man about to pioneer its involvement in modern music—the co-founder of ITOKA, Yihao Chen. Far from some supervillain, he was a very personable fellow who said all the right things, like “the goal of this AI technology is not to replace human creativity but to expand the territory or the boundaries of human creativity. In music, it looks at providing efficiencies.” And that it will “democratise the process” of making music, eliminating barriers to participation in an era where middleclassification is already rampant.

While a few mildly barbed comments about AI offering a shake-up to musicians who are “getting lazy” using the “same chords, similar melodies, similar rhythm” in the largely regurgitated world of pop, his message was that AI is just a creative catalyst of sorts and should help musicians earn more money. But then his ASMR-like tones soured, and he offered up the rather more startling remark: “Those people who are not willing to adapt and not willing to use this technology might suffer.” But how does that adaptation look, and what are the implications of using it? For this, even Chen can’t give a clear answer. This is where we now find ourselves in the music industry.

AI is a speeding juggernaut, and nobody knows if it has any breaks, steering, or who is behind the controls—already we’ve seen publications owned by media conglomerates covertly peddle AI companies they own, and fellow independents are few and far between anywhere in the music industry. This amplifies the scary position AI has placed us in.

(Credits: Far Out / Musicians’ Union)

“There are a number of potential problems posed by the use of AI as a more ‘generative’ tool where music is concerned,” Phil Kear of the Musicians’ Union tells me. “I say ‘potential’ as it is too early at this stage to see the real effect that AI will have on the music market. What we’ve seen so far are mostly novelty-type productions.”

The same can largely be said for its incursion in cinema. However, one of the many reasons actors and writers are currently on strike relates to this unknown. Where is the legislation, and how will their rights and wages be protected? In cinema, as in music, this is also being exacerbated by the current state of the industry as the confluence of streaming and the impacts of the pandemic continue to exercise an influence. So, can musicians ever follow suit and strike?

“At the moment, we are compelled to crystal ball gaze and try to figure out where AI will lead,” Kear continues, “and then try to generate a plan where human creators can still maintain a living alongside it. Unfortunately, that does mean I have to consider the worst-case scenario and plan accordingly, and that can lead to me often sounding like a Luddite technophobe doomsday predictor!”

“If we split the function of generative AI into two processes, there’s the input or ‘learning’ phase, and then there’s the ‘output’ phase,” he explains. “Both of these have a potential effect on human creators and creations. In order to produce AI-generated output, the AI needs to know how to write a song. To achieve this, it needs to be trained on huge datasets of existing copyright-protected recordings. Earlier this year, the UK Government announced a proposal to create a new copyright exception that would permit AI companies to conduct this training without any need to pay the creators or copyright owners in the recordings being used. There was significant protest from the industry.”

He continues: “The Government’s proposal would have allowed AI companies to produce output, using the previous productions of their competitors in the market, that would be in competition with those very productions they had used without any payment changing hands at all. Thankfully there was a hasty climb-down on that plan.” Nevertheless, its initial approval is indicative of the libertarian lawlessness currently applied to anything ‘online’. Moreover, the machines can hardly unlearn what has already been input.

“The output phase is even more complex,” he explains. “Let’s assume that the entire recorded output of the world to date has been used to train an AI, which then generates its own new composition. Does the output contain any element of the recordings used to train the AI? In your mind, you would say, ‘It has to; it only has the input to work from, so it must contain it’. However, the counterargument is, ‘Why is that any different to the human brain? As a composer, I have heard tens of thousands of compositions in my life – when I sit down to write a new song – how can it not be influenced by one or many of the already-existing compositions I’ve heard?’ So, there is a serious question as to whether the creators of the training recordings should be entitled to a share of the income generated from the new output recordings at all.” If you’re still with us, then that minefield symbolises the tangled issues.

However, the bottom line is part of the same battle that creators face at all levels of the industry already. As Kear explains: “The major concern for musicians is what share of the licence fee are they going to receive. Generally, in the current music industry model, they have usually transferred their rights to a rights holder (label, publisher). The rights holder not only has the right to decide if the music is licensed to AI but also will be the recipient of any fee. How much of any fee will be seen by the creator remains to be seen. If history is anything to go by, it will be minimal.”

So, it is likely that any potential profits from AI production will be near-enough nonexistent for the artists who are unknowingly/unwillingly part of the process. You might be happy to waver that if it was merely the case that your song was one of the billions that fed a machine to make some faddish facsimile, but the problem is that the subsequent output then shoulders in on your already squeezed pocket because AI will undoubtedly shrink the market for human artists.

Currently, one thing being promoted on the grassroots level as a way for new artists to work their way into the industry is SYNC licensing, whereby you either create or license sections of your music for stuff like a YouTubers jingle, a video game loading screen, a trailer for the new season of Made in Chelsea etc.. This both generates a small fee that you can then re-invest in your primary passion and helps you with some musical networking. As Chen told me, with ITOKA, anyone will be able to do this in five minutes by inputting something like ‘Rainy day, piano, cafe, sounds like Chopin but simpler’. So, as new organic markets appear, they are instantly being gobbled up by AI outputs operated by non-musicians.

As Kear states, this results in “less income and less opportunity for new acts trying to break through. Undoubtedly AI generated tracks are going to be cheaper and quicker to produce, and that is likely to have a depressing effect on payments that human creators can demand.” The life of a musician is somewhat precarious, as evidenced by the fact that a third of the UK music workforce was lost during Covid-19. “So, I can see a host of talented people giving up the industry altogether,” Kear grimly concludes. Chen, however, argues that AI’s streamlining of music will enable them to monetise their work easier, levying any losses and opening up new avenues.

(Credits: Far Out / Michael Sreenan)

This is the difficult possession where musicians find themselves. I spoke to two new up-and-coming artists in Benefits and A.S. Fanning, both of which have been fortunate enough to make a dent in the industry with two of the most acclaimed albums this year, and both of which exercised caution in their approach to AI. “I think we’ve got to be careful not to become too Luddite or fearful of this. It’d be properly obvious for any musician or creative to come out and say, ‘Nope, down tools, not having this, I’ve seen Terminator 2, fuck that’, but woah now, hang on,” Benefits’ Kingsley Chapman tells me.

This was seconded by Fanning, who remarked: “I try to take the attitude of accepting these sorts of changes rather than railing against them. It feels inevitable that AI is going to be a big part of the future, but it’s hard for me to see all the possible uses and ramifications from where we are now. Resisting these things can seem quite silly when seen in retrospect.” In essence, neither wants to be seen in the creative history books as the goon who said boo to Bob Dylan going electric.

The nature of music as a modern art form means that those creating it from the explorative position of a fledgling in the industry don’t want to short-sightedly close the door on something that is potentially expansive of the creative capacity. Moreover, something that could potentially be utilised to assist them in streamlining and monetising at a time when both of those elements are major hurdles. As Chapman muses: “Technology is kind of unstoppable, we’ll have to learn to adapt to it, I think. Lots of music is already enabled and produced by elements of AI anyway, so let’s not be too snobby and precious.” In other words, let’s treat it the same way The Beatles embraced Yamahas, which is wildly credited as innovative and enhancing, or the way Billie Eilish can now somehow sing over what is effectively an electronic big band in an emotive whisper.

When questioned on the creative impacts, Chen essentially said, ‘That if you can’t beat the machine, then c’est la vie’, but genuinely worthy human art will always eclipse it, and interestingly this is where Chapman and Fanning stand too. “What’s AI doing exactly? Generating music that sounds like other artists? No one complained when Britain was overrun by non-AI generated Oasis copycat bands for a couple of decades after Whats The Story,” Chapman wryly philosophised.

However, as Fanning highlights, the issue for musicians is more of an existential one in a legislative sense than a judgement day Armageddon for the arts. This is a point reflected by AI experts like Jaron Lanier, who have inferred that its presence in the arts is akin to saying that a car is a better runner than a human because it can go faster—it will always remain within our control by virtue of how it works, but that doesn’t mean the control won’t be crackpot. “If AI is able to imitate any well-known pop star or musician or band in a very believable way, then it could seriously limit anyone’s ability to earn a living,” Fanning comments—notably not that it could surpass.

(Credits: Far Out / Neil Hoare)

“I think if an artist has created something very distinctive, and AI is instantly able to imitate it, then it could seriously undermine an artist’s ability to monetise their work, which is obviously something that’s already extremely difficult for most artists,” he continues.

This is the battle that Kear has as a Union representative: “The Government are obviously keen to see AI adopted in the UK. There are so many advantages. The rest of the world are racing to make themselves the natural home for AI, and the UK needs to keep up. The current music industry needs a unified position on how it is going to work alongside AI and how creators and rights holders are going to fairly share income. This requires honesty from all parties, fairness, and trust to be gained where there has been little previously. It’s a fast-moving situation, and if this unity can’t be gained quickly, we could potentially see a compulsory licensing situation enforced by a frustrated government which could then see little control or income afforded to either rights holders or creators, and that really would be a doomsday scenario.”

And therein lies perhaps the biggest problem with AI. Even the most tech-embracing futurist would accept that some legislation needs to be in place to smooth the transition and offer guidance and equity. But who does that, and how, with a problem as tangled as AI, which isn’t stopping for anyone to work through the knots but rather accelerating into the unknown? No musician has either the time or resources to fully invest in this discussion, especially at the grassroots level – where, frankly, it is far more consequential – when they face a thousand other existential battles.

As is evidenced from the comments I’ve sourced, most musicians are happy to embrace AI in a musicological capacity, but they also have their concerns: where do they turn to with them? Chapman asks the same question about all the other loaded decks. “As for striking, who’s going to strike? Are Coldplay going to strike from playing huge shows? Blur? Springsteen? Doubtful. Why isn’t there talk of strikes because so many independent venues in this country are facing closure? Or in relation to streaming sites hoovering up all the royalties? Where’s the strike action to oppose support bands only getting a four-pack of Carling and some crisps? It’s an industry already battered by lockdowns and cost of living, not sure a performative strike over AI would help matters,” he says.

Fanning states: “I suppose I would consider striking if it came to it. At the moment, the whole AI thing is still sort of abstract in that I don’t really see people around me losing work to AI yet. Also, the idea of going on strike as a musician is a strange one as I would assume that nobody would give much of a shit if I went on strike, but if there was a broader movement towards it and people were really feeling the effects, then it would be something worth considering.”

(Credits: Far Out)

Both comments are indicative of the fact that musicians have seemingly seen a loss of autonomy over their work long before AI came along. Thus, for artists, it seems less like a Terminator coming to destroy a music utopia and more like yet another boot to the face while they’re down on their knees. This is rather despairingly symbolised by Chapman’s reflective comment: “What would a ‘music strike’ be? Not playing gigs? Not putting music online? Not writing? No one would give a shit, would they? If there could be a viable and effective way of protest, I’d support that, I suppose, but I can’t see how it would work in the same way that Hollywood actors all down tools.”

Kear has seen this notion more than most from his representative position in the Union, and he muses: “Do I see potential for strikes like the SAG one in the US? Music has shown itself historically to be an amazingly resilient and adaptive industry. It started off as a live industry only, successfully adopted recorded music, video, electronic instruments, downloads, streaming etc. Rather than rail against AI, I would hope that, yet again, the industry successfully manages to morph and subsume the new tech for its own ends. If there is a need for some kind of protest in order to get where we need to be, creators will be up for the fight!”

The history of music and tech is a promising one. Fittingly music is all about transcendence, so by its nature, it is up for the fight on both creative and external fronts. However, there is currently a severance within the industry that is hamstringing AI discussions, as musicians are too hindered by the damning battles of simply putting on a profitable gig or finding an affordable recording space to worry about a sci-fi plot. The hope is that AI can actually facilitate them positively, but that requires management and quick coalition.

In this process, it is more important than ever for transparency, honesty and independence to be paramount so that we can make AI another synth-block laughing stock in the history books of this beautiful art form and not the nail that sealed the coffin for young creatives already banging on the lid, and making enough beautiful music in the process to proudly prove the very substance of music is defined by defiance and advancement.

This is why Far Out remains proudly independent to support artists and institutions like the contributors above. And you can, too, by donating to causes like the Music Venue Trust. For more info on the Musicians Union, click here.

Source link

Comments are closed.