Perfect Fit Content (from elevator music to your AI DJ with Liz Pelly)

Hi, Paul.

Hey, Vass.

I want to play you a quick bit of music. Get your headphones on.

Okay.

I feel lighter because something went to my head there a bit, and I feel like I'm in an elevator going up to the 20th floor or something.

Maybe you are. Maybe you are. This is from the original Muzak catalog. It's one of the earliest forms of mood based music. It was designed not to be actively, closely listened to, but more to be felt, right? It was played in hotel lobbies, elevators as you said, department stores. The whole idea was to create a vibe.

Yeah, that kind of original vibe: new economy. You're shopping and you're floating with your hands on the cart. Your mind's wandering. You're grabbing items off the shelve, and you're zen and you just keep going because you're in the zone.

Exactly, and actually speaking of in the zone, Muzak pioneered something. It was called stimulus progression. Music would actually gradually become more energetic over time to increase things like worker productivity or customer movement. It was designed to get you hustling around that department store.

It's wild to think how far we've come with the music that we own or that we rent. We all used to own our music collection. You've probably been through it all, man. You probably had amazing records. Let's be honest, you threw on a cassette two. I used to burn CDs for my friends. We all have that one mixtape from our crush, maybe MP3s, and now our music libraries are largely digital.

Is yours? How do you listen to music?

Yeah, I've definitely gone digital, but I started out as a vinyl guy. As a teenager it was a pretty big deal to have a couple albums, and they were pretty expensive as I remember. And then suddenly there were cassettes, and 8-tracks shortly, and CDs. It was crazy. I've had all of them, and most of them then just you lose them. It was a bit sad to lose some of those gems and then rediscover them later digitally. But there was something about holding it in your hand as well, I have to say.

I believe you.

All of that shift when you stop and think about it and put it in perspective raises these really big questions about how streaming has changed both the economics of music, who earns money and how much, and our habits, our listening habits. Not to mention the way that we "discover" a new artist because that has changed, too.

Yeah, we're renters now largely. Spotify, Apple Music, YouTube, and we subscribe to things that are supposed to fit that. Here's your workout playlist. Here's your Zen playlist. Sometimes I don't want to be told the mood I should be in.

Fair enough.

But they can be helpful too to have those selections.

And so what about culture, this idea of discovery? I like to discover things as soon as I find old gems, but it doesn't jump out at you. Are we able to discover what we want? If you're looking to buy Canadian, does that really work in this kind of digital streaming? Are people finding what they want? Are they stumbling on the right things that you used to see in a shop window and stuff?

I don't know. I don't know how that's planned out.

Sometimes right now people are looking for music that helps them plug into their patriotism. I have seen news of Canadians changing their travel plans and how that's actually really hitting America, but I don't know if it's spilled over to hurting American artists, not that anyone would have that core intention.

As a side note, I do want to say my friend made these very tongue-in-cheek playlists. There's a series of them. They are on Spotify, they're called Big Shiny Tariffs, which is a plug on Big Shiny Tunes. We all had one of those CDs. They're Canadian bangers only and a patriotic algorithm hack.

Yeah, I love that label. The tariffs are coming. It's beginning of April. Big shiny tariffs.

That brings us right to the Spotify question, which is the focus today. I don't think they've put tariffs on Canada, but I'll have to check it. Are they a music company? Are they some kind of mega platform, an advertising company? What are they?

Yeah, if you were to take the lens of Canada's online streaming act, which is Bill C-11, the company would be characterized as a streamer, right? There's been a real policy push here in the Great White North to make Canadian content more discoverable, but how do you do that in an algorithm-driven world?

Yeah, you get squeezed out. It's the problem of the little piece in the puzzle. Is it swallowed by the rest? That's definitely happening in the way that these algorithms play out. We've got to promote some of those artists and Canadian content, and I'm not sure how that's going to evolve. I think there's going to be a push for it.

Something to keep our eyes on.

Well, to help us make sense of just how transformative Spotify has been as a company, whether it's an advertising company or a music company or something else entirely, we are speaking to music journalist Liz Pelly today. Liz is a writer and critic who explores the intersection of music, culture, and technology. She's best known for her incisive essays on streaming platforms, playlists, and the ways that Spotify is reshaping the music industry. Not just in terms of how we listen, but what gets made in the first place.

Her work has appeared in The Baffler, The Guardian, Pitchfork, and in New York Times. She's often cited for her deep dives into that mood-driven logic of algorithmic curation, and The Guardian called her new book Mood Machine "a savage indictment of Spotify."

Hi Liz, and welcome to Policy Prompt.

Hi. Thank you for having me.

We were wondering if you remember when you first heard about or saw or clicked on this thing called Spotify?

That's a great question and I wish that I had a better answer for it, but I don't really remember. I imagine it might have been while I was in the office of the newspaper I used to work at between 2011 and 2013. I think I have a memory of downloading Spotify onto my office computer there.

But it's not a vivid memory for me. Because to be honest, I wasn't really a huge streaming user when I started researching the company in 2016.

Paul, do you have any Spotify memories? I was feeling sorry for Paul because I think his music collection has been through too many evolutions, and maybe you lost track.

Yeah, it's true. I had vinyl albums, I had cassettes, I had all those kind of things, and they died. Each generation ended at some point, and then there was a break. I got into Spotify stuff around 2017, 2018, I would say. Set up some playlist and things like that. Found some oldies that I hadn't seen since cassette days or even albums, so that pulled me in. I'm not a huge user but a medium user.

Medium user. I remember that I was briefly using... I think it's Songza. You could put in your mood or it was music for getting ready to go out and then I was like, "Oh no, what a terrible precursor to all of this mood music stuff."

But I was playing it off my laptop. You know what I mean? I didn't have wireless speakers weren't as big. The sound was terrible. It was just like a laptop in your room while you get ready with a friend or something like that. Anyway, also not the most romantic or starry memory.

I mean, I do have a memory of in 2012 covering a music industry conference that happened in Boston where I lived at the time where an early Spotify executive was presenting about Spotify, and I do remember having really visceral reaction to... I'm not really sure if this is the path, like an immediate flag going off. But didn't not start seriously researching the company for another four or five years after that.

Right, Liz, so you've been covering music for some time, both as a personal level and business level. When did you see this as a platform that was going to be a game changer? When did that light bulb go off, whether you were using it or not?

Yeah, it's super interesting. I think that in 2016 there were a lot of conversations that were happening about the royalty model. There were a lot of people who had already started writing about and publicly discussing the financial impact of streaming on musicians, but it seemed like there wasn't as much public conversation happening around the way in streaming was impacting listeners.

I think just as someone interested in both music but also in media criticism, I started looking at Spotify and thinking about the user experience. That was where I came to it. It was just from these questions of when people open this app, what are they being shown? Why are they being shown it? What commercial considerations are shaping what people are being shown when they look at this app? In some ways those are the questions that I've continued following all this time.

There was a lot going on at that time as well, like smartphones. You mentioned the wireless speakers facet. There were a lot of technologies that were converging at that time to put this all in the spotlight.

Yeah, 2016 feels blurry to me. I feel like 2008 is an anchor year in a lot of tech and then anything before 2020 I feel like very collapsed in my mind, so I'm glad you're bringing that back.

But Liz, I'm glad you mentioned the consumer experience. Because I think if you don't remember your first time with Spotify, I think many people remember their first time downloading a song from the internet or streaming. It used to be, at least for our generation before it was super maybe normalized, it was something that felt and was transgressive or maybe pushing the boundaries of what we were supposed to be doing on the internet.

Maybe you could tell us a little bit more about what else was going on that prompted Spotify to form as a company, right? How did it capitalize on where Napster and maybe other file sharing music missed the mark and haven't had the longevity in our lives? We're also thinking of Pirate Bay in Sweden. Where does Spotify fit in that landscape and that kind of micro history?

Yeah, it's super interesting.

Spotify was created in 2006 in Stockholm by two men with backgrounds in the ad tech industries. At that point, the music business had been trying to recover from the impacts of the file sharing boom for years. The Napster years were in 1999 to 2001, and there's no shortage of different takes and perspectives on the impact of Napster and file sharing on the music business that are super interesting in their own right.

Yeah, the global recorded music business had been through the whole Napster saga and trying to figure out different ways to recover that in the mid-2000s. Something in the US that was starting to take hold as perhaps a new model for digital music was something like the iTunes store. It was this period of time where in order to access digital music this idea of the digital download store was championed or embraced by the music industry.

But something that is super interesting to the story of Spotify in particular is that that model never really took hold in Sweden in the same way. In fact, in the mid-2000s, piracy was still really popular in Sweden and also politicized in a different way. You mentioned you had the persistence of the Pirate Bay. Which whereas Napster was started by people who would go on to work in the tech industry and align themselves with tech giants, the Pirate Bay was started by people who grew out of this collective of punks and hackers and ravers who had created this activist lobbying group called Piratbyrån, which meant the Bureau of Piracy. They had a different ideological goal. In Sweden the whole idea of piracy was kind of more politically motivated than it was. You had a Pirate Party in Sweden.

(Clip):

Clara Elström is 22 and a Pirate Party activist. Like the other 18,000 Swedish party members, she says that all five sharing should be made legal, a core demand by the Pirate Party and one that earned it more than 7% in the last European election.

Their whole political agenda revolved around advancing the idea of file sharing and more conversations around is the impact of all this piracy like bad for society. There are more conversations around music as a public good, the internet as a public good. For the music business at that time, any sort of app or startup that would've emerged offering a product that was rooted in this idea of free music, Spotify's original business model was a library of free content supported by advertising, the music industry was super opposed to anything like that because they were trying to figure out how to recover from the era of free music.

But in Sweden, piracy was this really persistent issue. The music industry had started to view it as what people in the music business at the time referred to as a lost market. They were like, "No one in Sweden is ever going to buy music again, so might as well try out an app like this in this particular market." Of course through the negotiations with the major record labels that had to happen in order to license the content to make the Spotify streaming model possible, it didn't end up being the exact business model they had at first, free music plus ads. But in negotiating with the labels emerged this idea of the freemium model, where there's a free tier plus a subscription tier, and that's what we know now.

What made it compelling for the labels at that time, especially with all those pirate vibes and that characterization? Because that initial compliance is a fascinating element, too.

Yeah, totally.

I think that if you talk to people who worked in the music industry around that time, there's this idea that Spotify wasn't the first streaming service to emerge. Major labels had been interested in the idea of streaming as an alternative to file sharing for a long time. In fact in the early 2000s the major labels had tried to start their own streaming services and failed because their streaming services that they created were clunky. There was a lot of buffering time. The business model was weird and they never took hold.

There's also another streaming service in the US called Rhapsody that had launched a few years before Spotify. You also had Pandora. This idea of streaming wasn't invented by Spotify. It was something that the music industry had been pursuing or interested in. If you're to believe people who work in the music industry at the time, they would say things like, "Spotify just offered a more seamless product." Interesting hearing people talk about how Spotify arrived, it was like someone had finally created the streaming service that they'd been waiting for because of how fast it was on the technological level, that at that point there had been so many other streaming startups that had tried and failed.

There were a few different things they had going for them, I guess.

I remember the days of Napster and even Pirate Bay and that kind of frenzied moment where it was a crisis really, right? It did feel like a crisis.

Now Daniel Ek, who's the Spotify CEO, has had a massive influence on the streaming economy and how artists labels operate and he's still there, right? As you mentioned, he's the originator and he's still there as CEO. What's the world view that he's bringing to Spotify? What's his starting point?

Yeah, that's super interesting.

I mean, something that I write in the book is that one way of looking at the streaming story is that it could be viewed as the story of advertising men bringing the logic of their industries to music in new ways. ]I think that in a lot of ways, one, Daniel Ek and Martin Lorentzon, the co-founder who has arguably gotten more rich from Spotify than anyone else but doesn't get discussed as much in the public as much as Daniel Ek does, yeah, their backgrounds are in advertising. If you listen to Martin Lorentzon's early media appearances, he talks about how they had this idea for a product that was going to deliver a library of content funded by ads. But in the early days, they weren't even sure if it was going to be music. They had considered music. They considered video. They considered audio books or product search. They had a business model that involved advertising and they were looking for a traffic source.

I think that really stuck with me. I think the reality that in the early days, music was a traffic source for an advertising model is something that is very illuminating.

And then the other thing is that they're also publicly traded company that from my perspective doesn't seem to have a lot of goals or values when it comes to music and culture beyond the pursuit of profit to put it really just bluntly. I think that there's this sort of just disregard for... They're not music people.

Build a platform, see where it goes. If you build it, they will come, and it doesn't really matter who they are.

(Clip):

That's been the overhang issue on Spotify shares their secondary since their offerings, since they're listing, whatever was it, five years ago. The business doubled in revenue but gross margins never moved. The stock didn't work. Then when the gross margins moved, the stock worked. You got more evidence of it today. We continue to like the stock.

Well, they're also not in the hardware space. They're not doing anti-competitive things like tying or trying to... They're just even less creative it seems in terms of where else or what else.

But for people who use Spotify casually or maybe carelessly not understanding everything the company is doing, it could just feel like a very convenient music library, right? We see this all the time with tech conversations, convenience winning out, ease, loss of friction. Of course, your book argues that it's so much more than that.

How does the business actually function, right? Spotify, how should we think about it in terms of as a power structure, as a force shaping, reshaping the music economy? How can we move ourselves away from maybe appreciating a silly playlist that we came along to what it's representative of?

Yeah, so the company makes money from two ways: advertising and then selling subscriptions. I should say advertising even though it was the early goal of Daniel Ek and Martin Lorentzon, it is not a huge part of the business model. It's for years been less than 20% of revenue, and the vast majority of the money that the company makes is from selling subscriptions. Subscription is really important to Spotify, and it seems pretty clear that the company is really motivated by retaining users as subscribers. Like elsewhere across the platform economy, a lot of the strategic shifts seem to be built around hooking you on the platform in order to extend the amount of time you're spending on the platform every day in order to assert the value of the product in your everyday life with the seeming goal being to just make sure that people aren't canceling their subscriptions. Making sure that people are seeing Spotify as something that is a really integral part of their daily life.

There's lots of different ways in which that has played out over the years and the way in which pursuit of getting people hooked on the platform from early shifts away from... In the early days a platform like Spotify was more like a search bar, where if you wanted to listen to something, you would have to know more what album or artists you were searching for. There's a really strategic shift around 2012, 2013 where the product shifted. What Spotify was putting forth wasn't just this idea of this is a platform where you come and you could get access to the world of music. It started to become more about this is a platform where we are recommending and providing music for every moment in your life, the perfect soundtrack at the perfect moment.

It might seem just a branding exercise, but actually that shift ended up having a lot of consequences across music for both listeners and artists.

Paul has an AI DJ now or something, or some kind of infection on your computer. I don't know.

It popped up. I think it's real. I mean, I think it's intended, let's say. But it popped up maybe a year ago on my Spotify playlist streams of a DJ that was inserted. It says, "Hey, Paul, Saturday morning. I've got the perfect hits for you. I know you're in the mood for... I've been watching your playlist." It spews out stuff. I've used it a couple of times, but it's a bit eerie actually.

On AI, you talk about fake music or the ghost artist element and AI-generated tracks, anonymous filler, that kind of stuff. It's starting to become very hybrid and blended in the sense. Is that an area where we don't really know what's going on and there needs to be a bit of a framework around it, even regulation? Is this something that's just inevitable? Maybe you could talk a little bit about how that AI interface is evolving.

Yeah, I feel like there's two really distinct conversations. One is the way in which systems that might get referred to as AI impact the music that you are recommended on the discovery aspect of music. And then it's a whole other conversation related but distinct around AI-generated music and AI-generated content surfacing on these platforms.

You talk about the first bucket, take something like the AI DJ products that you just mentioned. I feel like that is a long time coming for Spotify, which Spotify really prides itself on not being new to the AI conversation, which is super interesting because it hasn't always been narrativized that way. Yes, since very early on in the history of the company, there has been an embrace of different forms of algorithmic recommendation, different forms of personalized recommendation, different ways in which machine learning has impacted recommendation.

In 2014, Spotify acquired this company called the Echo Nest, which is a big data for music company, which at that point had been powering algorithmic recommendation, machine learning-driven recommendation across many, many different music startups and companies and apps. But when they bought the Echo Nest, it signaled a new era in some ways in terms of Spotify's commitment to machine learning becoming an important part of the product. Over the years, we've seen all different examples of ways in which algorithmic recommendation has shaped what Spotify puts forth. It's like it's value proposition to listeners.

Whether it be something like the playlists Discover Weekly, the emergence of what they call Algo-torial Playlists, so playlists that are curated by editors but then sequenced to the individual user to create personalized recommendation of a playlist, the Spotify homepage since 2016 has been populated with things like Daily Mix, which is personalized algorithmic playlist, Release Radar. There's the whole Wrapped campaign every year, which depends on interpreting your user data in different ways. Spotify takes a lot of pride in being a data-driven company and in their own marketing materials.

We'll talk about how that's from their perspective one of the things that people like about Spotify is this sense of personalization. It's really interesting looking at when they shifted into starting to call that AI, so AI DJ is a recommendation product. You press this one button and it gives you different blocks of algorithmically generated recommendations that are built on pre-existing algorithmic recommendation models that they had been using for a really long time. And then there's a generative AI voice that pops up in between these different blocks to announce what's going to be recommended next.

But in a lot of ways, it's a repackaging of things that had already existed on the platform for a long time.

Policy Prompt is produced by the Centre for International Governance Innovation. CG is a non-partisan think tank based in Waterloo, Canada with an international network of fellows, experts and contributors. CG tackles the governance challenges and opportunities of data and digital technologies, including AI and their impact on the economy, security, democracy, and ultimately our societies. Learn more at cgonline.org.

All sorts of government officials, whether they're elected or bureaucrats, may have the app. You know what I mean? It may be a soundtrack to some of their governance work. You do note that governments have largely ignored Spotify's dominance or growing dominance, right? It's not the only streamer, but it's leads the pack. Why do you think our regulators from a policy side are overlooking a company like this? Is it because it just doesn't fit neatly into traditional regulatory categories, right? Is it music and culture? Is it ad tech as you're saying? Is it a media company? What are we missing because there's so many fascinating things to pay attention to here?

Yeah, it's a super interesting question.

I mean, I think part of it might be related to just the broader reasons why also in some ways consumers have accepted the type of surveillance that is required in order to produce a product like Spotify. I think that there is this sort of general sense of either resignation or acceptance. When people talk about streaming surveillance for example, something that you'll hear a lot is like, "Oh yeah, they're collecting all of my data, but it's just music." "Oh, but it's just in order to recommend me music." There's a sort of hesitance.

They know everything anyway. That comes up too, right?

Yeah, totally.

I mean, I don't agree with it.

Yeah, I mean, same. Yeah, I feel like there's this kind of hesitance to consider it outside of the context of it's just music, but obviously there's all sorts of concerns from a surveillance perspective and the use of data, sale of data, involvement in the data marketplace. I think it could be just a general sense of feeling like there's bigger policy concerns than this thing that's just music.

But I also think that what you just said is also correct, that maybe there's a confusion about the angle with which to consider this. Because like you said there's in some ways when we talk about the issues with music streaming services, we're talking about the issues with corporate consolidation and the corporate conglomerates that control the recorded music market more broadly. It's like at what extent are a lot of these issues that come up with streaming actually connected to the outsized power that Sony Universal and Warner have over the trajectory of recorded music? Is there an antitrust conversation with regard to the influence of the majors that needs to be contended with first?

Then there's other conversations around specifically practices of platforms themselves. One of the things that I track in my book is these different cost saving initiatives that a company like Spotify has employed, and I think that they raise different issues around digital payola. That could be another angle through which regulators might approach this issue.

And then obviously there's also issues having to do with the royalty system that relate to both of the previous topics.

Well, let's get into that a bit.

But before that, let's just note that governments around the world are very worried about these services that get literally inside your head, right? This is one of many actually, but isn't there an element of this is what people want as well? Their lifestyle has changed. They're busy doing things. They're driving. They like to listen to things that are offline. It's giving people what they want very much. The culture has changed around music and listening and things. And so to a certain degree it's responding to the market, but I think people are quite unaware of what's going on behind the scenes, both as you say in terms of the royalty regimes. How do you make it as a small artist? Are you listening to the back of the record, the flip side of the record, the old vinyl record? Do those things get play? All that stuff's been washed away.

People maybe just assume it's fine, but maybe you could say a bit more about just how the royalties work. How do you break out as a small artist in this kind of space through Spotify?

Yeah, and I think that's a really good point too because something that comes up a lot talking about the impact of streaming is this idea that, well, the average user doesn't care. But I actually think that a lot of the issue is rooted more in the reality that people don't know or don't have the information or context. I think that when you do start highlighting some of these issues, what gets revealed is that more people than you'd think actually are concerned with the impact of these services on listeners and on artists.

But to answer your question about the royalty system, I think that in a lot of ways, the ways in which royalties are paid by streaming services, this is a root cause of some of the various different ways in which our ideas of value and music has been shifted.

There's this common figure that gets tossed around in the streaming era that musicians make on average like .0035 cents per stream.

I've seen that, yeah.

Yeah, it's a popular headline or metric or figure that people use, but it is not a super accurate way of thinking about streaming royalties because streaming services don't pay per stream, and they also don't pay artists directly.

To get your music onto a streaming service, an artist has to go through a rights holder or some sort of entity that has a contract with Spotify. If you're on a major record label, if you're on Sony Universal or Warner, your music will get service to Spotify through those companies that directly distribute their music to streaming services. If you're on a big independent record label, there's this organization called Merlin Group that represents the bigger independent record labels. They have a deal with Spotify and other streaming services. And then there's also these sort of self-serve companies like DistroKid and CD Baby that more like DIY artists who are like, if you recorded a song tomorrow and wanted to get it onto Spotify, you could go through DistroKid or CD Baby. They have a contract with Spotify.

The payments are made to those rights holders based on a pro rata system. It's this complicated revenue share model, where it's not that you're getting paid on a per stream basis, it's that the rights holders' catalog is being considered according to stream share. If you're Universal Music Group and your catalog accounts for 20% of total listening on the platform during a royalty period, then you would be paid 20% of the royalty pool according to the terms of your contract. Each contract specifically dictates what the eligible royalty pool even is.

There's so many levels of abstraction here in terms of knowing who's getting paid what. And then the rights holders are given these listening reports that they can take that, say, how many streams each artist got. They can take their slice of the pie, and then they turn to their artists and they pay artists their share of their slice of the pie based on their contracts with their artists. There's multiple levels to it.

Amazing. Lots of legal layers here. I see a lot of lawyers involved in these processes, which sounds very American in a way I must say, too.

Well, yeah. It's hard to imagine, I think. I think people may expect that their dollars are going directly into something that is distributed based more on their personal listenership than understanding this big massive pool and all the skimming that's happening.

You mentioned a concept payola, which actually might be useful to define for the listener, and then maybe we can follow up on it by asking about the discovery mode model, right? Spotify's newish system, where artists make a very particular bargain. They are accepting lower royalties, lower than the low we've already described, in exchange for greater algorithmic visibility that dangling dangling of Spotify's power. It has been compared to modern day payola.

You mentioned it. Do you see that system as payola? As I said, what the heck is payola anyway?

Payola has this really specific legal definition or context, at least in the US, where it refers to this practice from the era of commercial radio. The radio station bringing a pile of cash and sliding it under the door in exchange for radio airtime in exchange for radio play. It really specifically refers to an exchange of compensation in exchange for airtime. It's illegal. It was made illegal in these famous payola hearings.

When we think of the types of commercial considerations that shape streaming curation or commercial considerations that shape the promotion of certain content by certain platforms online, it's better to think of it as a payola-like practice and not necessarily specifically a reproduction of payola. But I think that there are many similarities that these practices have with the reasons why people historically have been concerned with payola, which is that...

For example, something that in the book I refer to as a payola-like practice of the streaming era is this program that Spotify runs called Discovery Mode, where a rights holder can accept a 30% royalty reduction in exchange for algorithmic promotion on things like Daily Mix, Radio, and autoplay on Spotify. It's not exactly the same as payola, but there are some similarities. One being that when a listener is recommended a track because of its inclusion in the Discovery Mode program, it's not made clear to them. I think that that's the way in which there's the possibility of a commercial consideration. Inflating a sense of popularity for a track, a listener being recommended something, it not being clear why they're being recommended it is definitely made, I think, a similarity there.

Historically, how has this worked? How are artists supported?

You talk about examples in France and more recently in Ireland about living wages and how that's gone over time. And then of course in the United States there's a Living Wage Musicians Act and there are in other countries as well, so it's a very hot discussion policy-wise. But there's also some history here of what things have worked in the past or recent experiments perhaps in Europe. You could talk about a couple of those.

Yeah, so I think something that I write about towards the end of the book is that I think that the failures of the streaming model for independent artists do expose these deeper issues around at least in the United States where I'm writing from, there being barely any public funding for music at all. Maybe it's not even just a matter of the reality that streaming services need to be reformed, that artists need to work on building alternative models for supporting culture, but also what does this say about how necessary it is for us to think of different ways to publicly fund music? Or what does this say about the need for more public support for music in the arts more broadly?

I mean, I think that something that people like to point to is there was a really short history in America in the New Deal years where there was this brief moment of the government creating different publicly-funded music programs. But for the most part, there's not a lot of public funding for music, especially outside of classical or music that has more institutional support generally.

FDR, New Deal, 1930s, is that what you mean when you say New Deal going back?

Yeah. There was something in the United States called the Federal Music Project, and I think during the early COVID years that was being discussed a lot in the United States because people were talking about how badly in the US we need something like this, some sort of federal program that could employ musicians or fund public performances or provide teaching jobs for musicians, et cetera.

There's a brief historical precedent for this in the US, but I found it to be more instructive to look around the world at current examples of countries where there are systems in place for supporting the arts. Yes, I pointed to in France the type of specific unemployment scheme that exists for artists to perform a specific number of public performance hours every year, or in Ireland this basic income program for the arts that was piloted during COVID.

Yeah, if you talk to artists from both of those places, they will tell you that these systems are imperfect. But I think they do provide examples of how, especially when we're talking about music, it's even a little bit of public funding or public support goes a long way for people who are navigating such a precarious industry.

Yeah, it could be interesting and important in the interim.

Before we let you go, I want to squeeze in three questions. I'm going to make this next one a two in one, which is highly illegal. How do you think policymakers could better engage musicians and even listeners as they try to or continue to shape policies that affect the future of streaming and music creation?

And then I also wondered if you were designing, if you had your druthers, if you were designing a more artist-friendly music streaming economy or music economy, what would it look like maybe beyond baseline funding for creative workers? What is the role of the state in helping people realize that?

Yeah.

Well, I think that something that I tried to emphasize in my book, something that is a little bit unique about it, is that in addition to interviewing a lot of people who work in the music business or people who work in policy for example, a lot of musicians are interviewed. I definitely feel like including the voices of musicians and the research process.

In the UK, something I cite is this Musicians' Census that the UK Musicians Union has started doing. I definitely feel like there is a need for more funding or support for those types of musician census projects so that we have more data directly from musicians about how their lives are being impacted by the shifting dynamics of the music business. I think that could be a really helpful model for policymakers to look to your support or try to partner with musicians unions or different advocacy groups to have more information. And then anything that could be done to bring musicians and music workers into the fold would be, I think, super helpful.

In terms of thinking about policy, I also think something that there's these baseline things that I think need to happen in order to make informed policy decisions or even to know what questions need to be asked. One of these is more information about the experiences of musicians.

Another thing is that the secrecy that exists around how streaming services operate I think is a really a big hurdle in knowing what the right questions to be asking at a policy level even are. There's some research I point to in my book that was done a few years ago by this DC-based think tank called Public Knowledge. The name of the research is Streaming in the Dark, and it highlights the role that NDAs and black box contracts are currently playing in creating obstacles to further policy interventions. What are some ways in which policy makers could require oversight of those contracts in order to actually take a closer view of what are the terms that are actually dictating these contracts and these systems that impact artists and listeners? I think that that also is worth looking at and thinking about in policy conversations.

Then the second part of your question, what model would I look to, I mean, there's definitely a reason why towards the end of my book I spend 10 pages talking about public libraries. Not just the role of public libraries and providing access to library card holders to digital music, but also there has become a trend across the US and Canada of public libraries hosting these local music streaming projects that are somewhere between offering a music streaming service of local music and also fulfilling this role of archiving local music.

Really interesting, it actually is a public library in Edmonton, Canada that I interviewed a librarian from who put forth this idea of the local public library-funded streaming program as not just an alternative to Spotify, but as a digital public space, where people involved in local music communities are invited in to have a say in helping shape the digital tools that they rely on in their own communities. I think it's just such an instructive model and also tells us so much about the problems with streaming also.

These aren't technological problems. It's not that there's an issue with the idea of having some music held on a central server and using the internet to access it. These are issues of power and industry practices informing the technology.

And infrastructure. Well, go Edmonton.

Liz, thank you so much for your research, for your advocacy, for your relentlessness, and for breaking down Mood Machine for us.

Thank you so much for having me.

Thanks, Liz.

All the best.

Okay, so perfect fit content, PFC is what Liz calls it in the book, I feel like that phrase is going to live in my brain for a little bit now.

Yeah, right? The music industry version of you might also like, and what about this? It's on steroids and it's just coming at you.

I've got something right here I'm actually just could put on is on my Spotify this feature emerged that's like your DJ, and I never put what it's going to say. It seems to spit out Pink Floyd all the time. Let's see what it does right now. I'm literally going to put it on.

AI DJ:

Hey, what's happening? DJ X got your usual selection ready to go. I know what you listen to. I see Scorpions there, so I'm going to be here every day playing those artists you got on rotation going back into your history for songs you used to love. I'm always on the lookout for new stuff.

There you go. Scorpions is what it picked for me right now. Actually, the funny thing is, the very first rock concert I ever saw was Scorpions when I was a teenager.

I mean, that's creepy that your AI DJ might know that or that he calls you a scorpion. He seems quite familiar with you.

It's funny how you don't even notice the synthetic content sometimes until someone points it out. And then you wonder, "Have I been listening to ghost artists all week? What is behind that chatbot that's my DJ?"

Yeah, totally. And so there's a genius in it of just the ability of the algorithm, and then it's super creepy because it's right in there, and they're calming you down and like... I don't know what they're feeding me, right? Maybe they're pushing me in other ways, too.

Which does bring us to that concept of streaming fraud too, right? The idea that not only can music itself be fake or artists be fake in terms of how they're presented to us, but that there are artificial listening, right? Artificial boosts from bots or fan clubs trying to juice somebody's statistics on a platform feels like it's becoming just part of the game now. Maybe that should be a bigger deal.

Yeah, you put your finger right on it, right? Who's the referee? Who's the cop on this because listeners are just vibing and the platforms are happy. No one's blowing the whistle when things go offside.

Right. Listeners are just vibing.

The other thing that I feel like no one really wants to talk about, and Liz gestures at this in her book, is remember silence? Just sitting in a car without a podcast or walking around to pick up the mail without listening to a playlist?

Yeah. I mean, I don't know what to say on this one, but I love silence still. When I go cross-country skiing or something and the branches aren't even moving because it's the absolute still, I have still not introduced podcasts to those moments.

I hate a noisy branch.

Noisy branches, but it's coming everywhere. I feel like I have to have a podcast or something on almost now. You're right. We're not supposed to be silent.

It partially also plays into those self-optimization pressures too, right? Where it's like, "Oh, I am cleaning the apartment, but I could also be listening to something," or learning something or just zoning out. Maybe our potential aversion to unplugging isn't personal. Maybe it's more systemic. There's just this whole economy built around keeping us tuned in, keeping us listening, even if it's songs are also made by nobody.

Yeah, so totally true. Thanks to Liz Pelly today for helping us see the algorithmic forest despite those playlist-shaped trees all around us.

I appreciate her book. Very much loved our conversation. Listeners, try a little silence if you dare.

Policy Prompt is produced by me, Vass Bednar, and Paul Samson. Tim Lewis and Mel Weersma are our technical producers. Background research is contributed by Reanne Cayenne, brand design by Abhilasha Dewan, and creative direction from Som Tsoi. The original theme music is by Josh Snethlage, and special thanks to creative consultant Ken Ogasawara.

Please subscribe and rate Policy Prompt wherever you listen to podcasts, and stay tuned for future episodes.