Is Streaming Broken? The Big AI Fraud Indictment Raises Serious Questions

You can’t say no one’s getting rich from streaming. In an indictment unsealed in early September, federal prosecutors charged musician Michael Smith with fraud and conspiracy in a scheme in which he used AI-generated songs streamed by bots to rake in $10 million in royalties. He allegedly received royalties for hundreds of thousands of songs, at least hundreds of which listed as co-writer the CEO of the AI company Boomy, which had received investment from Warner Music Group. (The CEO, Alex Mitchell, has not been charged with any crime.) 

This is the first criminal case for streaming fraud in the U.S., and its size may make it an outlier. But the frightening ease of creating so many AI songs and using bots to generate royalties with them shows how vulnerable the streaming ecosystem really is. This isn’t news to some executives, but it should come as a wake-up call to the industry as a whole. And it shows how the subscription streaming business model with pro-rata royalty distribution that now powers the recorded music industry is broken — not beyond repair, but certainly to the point where serious changes need to be made.

One great thing about streaming music platforms, like the internet in general, is how open they are — anyone can upload music, just like anyone can make a TikTok video or write a blog. But that also means that these platforms are vulnerable to fraud, manipulation and undesirable content that erodes the value of the overall experience. (I don’t mean things I don’t like — I mean spam and attempts to manipulate people.) And while the pluses and minuses of this openness are impossible to calculate, there’s a sense in the industry and among creators that this has gradually become less of a feature and more of a bug. 

At this point, more than 100,000 new tracks are uploaded to streaming services daily. And while some of this reflects an inspiring explosion of amateur creativity, some of it is, sometimes literally, noise (not the artistic kind). Millions of those tracks are never heard, so they provide no consumer value — they just clutter up streaming service interfaces — while others are streamed a few times a year. From the point of view of some rightsholders, part of the solution may lie in a system of “artist-centric” royalties that privileges more popular artists and tracks. Even if this can be done fairly, though, this only addresses the financial issue — it does nothing for the user experience.

For users, finding the song they want can be like looking for “Silver Threads and Golden Needles” in a fast-expanding haystack. A search for that song on Apple Music brings up five listings for the same Linda Ronstadt recording, several listings of what seems to be another Ronstadt recording, and multiple versions of a few other performances. In this case, they all seem to be professional recordings, but how many of the listings are for the same one? It’s far from obvious. 

From the perspective of major labels and most indies, the problems with streaming are all about making sure consumers can filter “professional music” from tracks uploaded by amateur creators — bar bands and hobbyists. But that prioritizes sellers over consumers. The truth is that the streaming business is broken in a number of ways. The big streaming services are very effective at steering users to big new releases and mainstream pop and hip-hop, which is one reason why major labels like them so much. But they don’t do a great job of serving consumers who are not that interested in new mainstream music or old favorites. And rightsholders aren’t exactly pushing for change here. From their perspective, under the current pro-rata royalty system, it makes economic sense to focus on the mostly young users who spend hours a day streaming music. Those who listen less, who tend to be older, are literally worth less.

It shows. If you’re interested in cool new rock bands — and a substantial number of people still seem to be — the streaming experience just isn’t as good. Algorithmic recommendations aren’t great. Less popular genres aren’t served well, either. If you search for John Coltrane — hardly an obscure artist — Spotify offers icons for John Coltrane, John Coltrane & Johnny Hartman, the John Coltrane Quartet, the John Coltrane Quintet, the John Coltrane Trio and two for the John Coltrane Sextet, plus some others. It’s hard to know what this means from an accounting perspective — one entry for the Sextet has 928 monthly listeners and the other has none. If you want to listen to John Coltrane, though, it’s not a great experience.  

What does this have to do with streaming fraud? Not much — but also everything. If the goal of streaming services is to offer as much music as possible, they’re kicking ass. But most consumers would prefer an experience that’s easier to navigate. This ought to mean less music, with a limit on what can be uploaded, which some services already have; the sheer amount of music Smith had online ought to have suggested a problem, and it seems to have done so after some time. It should mean rethinking the pro-rata royalty system to make everyone’s listening habits generate money for their favorite artists. And it needs to mean spending some money to make streaming services look more like a record store and less like a swap-meet table. 

These ideas may not be popular — streaming services don’t want the burden or expense of curating what they offer, and most of the labels so eager to fight fraud also fear the loss of the pro-rata system that disproportionately benefits their biggest artists. (In this industry, one illegitimate play for one song is fraud but a system that pays unpopular artists less is a business model.) But the industry needs to think about what consumers want — easy ways to find the song they want, music discovery that works in different genres, and a royalty system that benefits the artists they listen to. Shouldn’t they get it? 

Related Images: