Podcasts
Podcast: A Panel Discussion about Music Licensing, Royalties, and AI
Read Time: 11 minsMusic royalties can be a complicated process. The process got a little bit more complicated with the introduction of digital and streaming revenue, but attorneys, producers, managers, and artists all want to know how they get paid. I’ve got a couple of great guests with me today to talk about how music publishing is paid, the royalty streams, and about the introduction of artificial intelligence (AI) and what that might mean to the music industry’s royalty flow.
Hillel Frankel: Hello everyone; I’m Hillel Frankel. I’m an attorney at McGlinchey Stafford. I’m here in the Nashville office. I work primarily with music and arts, and entertainment companies, and we’re joined by two distinguished guests. Chris Allen is the Head of Artist Relations at Stem Distribution and Royalty Services, and we also have Nathan Osher, who is the Assistant General Counsel of Publisher Relations at the Mechanical Licensing Collective (MLC). We are going to do a reprise of the publishing panel that we did at Jazz Fest a few weeks ago. Let me start with Chris. Chris, tell us a little bit about what you do related to Stem Distribution and also about your history in publishing and sync licensing.
Chris Allen: I have been in the business in various capacities for the past 20 years. I primarily started out on the artist management side, and I would say in the last ten years, I’ve been much more focused on music publishing, publishing administration. Then in the last five years, really an emphasis on digital distribution, working with a couple of different digital distributors that deliver music to digital service providers (DSPs) worldwide.
Hillel Frankel: Sounds good. And Nathan, maybe just a brief introduction regarding the Mechanical Licensing Collective and what the MLC does, and who they collect from. And then I’ll shoot a couple of questions at you guys following the intro.
Nathan Osher: Mechanical Licensing Collective, which I’ll refer to as the MLC, we are a nonprofit organization that was designated by the United States Copyright Office pursuant to the Music Modernization Act of 2018. We are the exclusive administrator of the compulsory blanket license that is relied on by DSPs, which are digital service providers, or digital music providers (DMPs) as they’re referred to in the Music Modernization Act. They’re relied on by the DMPs for the distribution of digital streams and downloads of recordings of musical compositions within the United States for private use. The MLC we are, as I said, we’re the exclusive administrator of the license.
We are also charged with creating and maintaining a database of musical works, which is pretty much a complete database of all musical works, not only distributed in the U.S., but also contributed from outside organizations and Collective Management Organizations (CMOs). They also register their works in our database if the works are distributed in the United States. And so far, we’ve done very well. We’ve got over 28,000 members. Those are either self-administered songwriters, music publishers, administrators, or CMOs, and over 31 million registered works. We launched on January 1st, 2021 and to date, as of our May distribution, it’s just over 24 distributions. We’ve distributed, we’ve just passed the one billion dollar mark, so we’ve distributed over a billion dollars.
Hillel Frankel: A billion dollars to publishers and songwriters?
Nathan Osher: Publishers and songwriters, correct. This is for the digital use of their recordings of their musical compositions. You’ll get into the copyright composition and sound recording.
Hillel Frankel: Let me take a minute then to talk a little bit about musical product and what copyrights are related to musical recordings. There’s actually two copyrights. One is for the sound recording, the actual recording of the music, and one is for the composition. And if you know music, you know that you can have a song that’s composed or written by any great artist or great songwriter. And that song could be covered by multiple people. It could be covered by the artist themselves who wrote the song, but also by multiple people. So for that reason, very specifically, there are two copyrights and two royalty streams from which music entities and music writers and artists derive royalties.
Generally, the copyright, the sound recording copyright side, is affiliated with the record label side, and the composition side is affiliated with the music publisher. Generally, in the history of music business, we’ve had two separate entities collecting royalties for artists and songwriters. As things have developed digitally, there are multiple revenue streams and there are multiple ways to collect. So I’m going to have Nathan follow up, and then I’ll ask Chris, but with regard to the publishing, you were talking about mechanical royalties collected by the MLC. How does that generally break down? What are the royalties that are collected, and what percentage, if you can actually say the percentage, might go out per stream?
Nathan Osher: Per digital streams and downloads, DSPs are required to have two types of licenses. One is a public performance license, which they will get from the performance rights organizations, the PROs, and that’s the American Society of Composers, Authors and Publishers (ASCAP), Brodcast Music, Inc. (BMI), Global Music Rights (GMR), the Society of European Stage Actors and Composers (SESAC), those are the big four. And then there’s a mechanical royalty. The mechanical royalties are paid to the MLC, and the mechanical royalties are based on a rate. Recently, the court came back on the Phonorecords III remand, so we now have Phonorecords III rates. It has to go through a process of 60 days for any type of comments. And then, once the rates are final, the DSPs will pay, within six months, Phonorecords III rates. So right now, there’s a Phonorecords I, II, III, and IV rate, and those are based on percentages. And it’s a complex computation based on the type of service the DSP offers. But in the end game, it’s a royalty rate that the MLC collects and pays to the members. Who are, again, the songwriters, if they’re self-administered. If they’re represented by a music publisher or publishing administrator, we’ll pay it to their representative or to a CMO, which is the collective management organization, and they will in turn pay out to the songwriters.
Hillel Frankel: And this rate was determined after a series of litigations by the copyright royalty board. Could you name the players in that litigation so that we know where the number has come from and these various rates?
Nathan Osher: Well, it’s really the DSPs and the representatives of the music side of the business, the publishers, the labels are involved. So it’s across-the-board business, music business.
Hillel Frankel: And they’ve come up with a negotiated rate. This current rate you just described is a 40% increase from the previous rate?
Nathan Osher: Yes, it’s a significant jump, which is a very good thing for the songwriters and the publishers.
Hillel Frankel: That’s great. So, Chris, there’s also another way that publishers and songwriters make money, and that has to do with sync licensing or synchronization licensing whereby music is affiliated, synced to a film or TV show or advertising. That has become a significant revenue stream for a lot of songwriters and publishers. Can you tell us a little bit about that synchronization licensing, and give us an example of something that you might have placed or some song that you are affiliated with that got a good placement?
Chris Allen: Anytime sound recording is synced up with a visual and it’s broadcast, whether in a TV show, a film, or a commercial, the entity using the music has to exercise a sync license, and that’s when they usually contact the publisher and the label. So there are two sides to that payout. Initially, there’s a payment that goes to the sound recording side and then a payment that goes to the publisher’s side. Usually, they’re equal and follow along with each other. And in addition to that, once the show is broadcast, usually not with the exception of a commercial, but if it’s a television show or a movie and it gets played around the world, then that’s when your affiliated PRO that as a songwriter or a publisher, they collect that public performance of the repeated airings of the show, which thus your music gets played more.
Hillel Frankel: Usually, an advanced payment or fee payment, and then a backend payment, which is a royalty for the performances, depending on where it’s performed: TV, cable, around the world, film, and also film, when the film is on TV and cable.
Chris Allen: Yes, and depending on the show, and maybe if you had a long-standing show like Seinfeld or The Simpsons, something like that, you can continue to collect residual income for many, many years.
Hillel Frankel: It’s a good revenue support for songwriters. Moving forward, just to look at where things are going. We recently have heard about some AI cases, some situations where artists, possibly an AI-generated artist, has created a song that’s a sound-alike or uses the vocal tones of famous artists and put those songs up through the DSPs, through the digital service providers.
As we close our summary of publishing and rights with regard to copyright, I wanted to ask you guys what you think of the situation? What would happen if a similar song was run through your company? How would you be able to edit it, or what would be a copyright solution? I’ll start with Nathan on this. If the MLC determines that there’s a song has been published or released and it’s getting significant streaming plays, but it possibly is the creation of AI and not necessarily an artist or publisher, how would the MLC deal with royalties? I know you have a music litigation background; what would be your opinion about something like the recent Drake and The Weeknd AI-created song that was up for a little while and then taken down?
Nathan Osher: From the publishing side, you have to look at the copyright in the musical composition. AI is very interesting because if you have a sound-alike recording of an existing musical composition, I would call that a cover recording. And as far as the composition copyright, the copyrighted musical composition embodied in that recording, the original publishers, the songwriters that wrote it, and are the publishing entities, would continue to own just like a regular cover situation. If you or I covered, using your analysis of a Beatles tune, the Beatles writers would still own the copyrighted composition, and we would pay the royalties to those owners of the composition copyright, or their publishers.
When you’ve got a work that is created by AI, meaning it’s not a cover recording, but it’s a new work that’s created by AI, you have to start looking into how much of it is created by AI. If it’s under the Copyright Act, a musical work is copyrightable if there is some element of human contribution, a human has to create the work, the musical work. So if you’ve got a work that’s completely created by AI, then I would say that’s not copyrightable.
It’s still very new in the music business; we don’t really have case law. We have guidance from the copyright office. There are a couple of cases or applications that they’ve reviewed and published their decisions. If you have a work that’s completely created by AI, it would not be copyrightable. However, you can have a work that uses uncopyrightable elements, but a human has come in and arranged those elements, and that would be copyrightable. So what we really need is some type of rule of thumb on how much human contribution is required for a work to be protected. And then we have those same guidelines like in regular music cases, where you’ve got a songwriter that uses unprotectable elements and arranges them to create a protected work. That arrangement would be protected.
Hillel Frankel: If the arrangement is created by a person, there is some copyrightable element there that might be owned. The interesting thing about The Weeknd and Drake AI song that went up, I believe the way they were able to get it to come down was there was a rights of publicity in the voice. The artist owns the right to their own vocal sound, their own vocal style, but also I believe there’s also a tag. There was a copyrightable element that was included in that song, and they were able to say, we have rights to that tag, the producer tag or whatever element was sampled in there. And that had to come down because Universal had exclusive rights to that, to distribute that piece of music, and there was no license for that distribution.
So Chris, kind of along the same lines, you work for a digital distribution company. Stem services the DSPs with music. How would it work at your company if there was a composition that was submitted that you thought was created by AI or maybe was somewhat of a sound alike? Something created not by humans, but by AI artificial intelligence.
Chris Allen: It’s going to be different for each distribution company. Stem is a very inclusive company in terms of, who we allow to use the service, but there are many do-it-yourself platforms out there, and I think that’s where the exposure’s going to come from. I just saw stats recently that there are now 120,000 new songs a day being distributed to all the platforms. So that’s close to a million new songs a week. Because Stem and then I think as you look at record labels in particular, there’s a lot of eyes, ears, and hands touch and are involved in the setup of the song and you know, when it comes out. So I don’t think it’s going to be one of us that is the culprit, so to speak, of distributing an AI song, but I do think it’s going to be a major issue.
Chris Allen: We also get rated by Apple, Spotify, Amazon, etc. in terms of making sure that our content is delivered within all their guidelines, and there aren’t any infringements on the stuff that we’re delivering. Whereas the open forum platforms, the DIY platforms are a little more loose in terms of what they deliver to the DSPs. It is going to be an issue when somebody sounds like Adele or somebody sounds like Sam Smith or a distinct superstar’s voice, and it’s not one of their songs, and it does get delivered. The good thing I think about the way the system works is once you deliver a song, there’s about a two to three-month lag between the song actually being out in public and when the money gets paid to the person that delivered it. So there is some time, especially if the song does start to sort of raise its hand above the noise of all the other songs that were released. I think people will then notice that it was pirated, AI, or something that’s questionable about the infringement. So I think it’s going to happen more often, and like I said, it’s inevitable, but I think there are enough people that are paying attention. And also, because of the lag on when the money gets distributed, there’s a way to catch it before anybody makes money from somebody else.
Hillel Frankel: Yeah, it’s interesting. And they did catch the other one, I think, within 24 hours. They had pulled that Weeknd at Drake copy song down. But this is great. It’s an interesting discussion, and I know during the panel at Jazz Fest, we never got to this part, so I got to sneak it in here on the podcast. I want to thank Chris and Nathan for taking the time and being with us this afternoon. Thanks, guys. Hope to see you soon.
Nathan Osher: Thanks for having us. If you want to find out more about the MLC, the website’s very informative, it’s themlc.com.
Chris Allen: Yes. Stem is www.stem.is
Hillel Frankel: Great. Thank you both of you. Thank you very much, guys. Bye-Bye.
Subscribe wherever you listen to podcasts: