Recording Academy Warns AI Music Threatens Creators (Full Transcript)

Panos of the Recording Academy discusses streaming royalty inequities, dual copyrights, and why AI-generated music needs transparency, laws, and creator control.
Download Transcript (DOCX)
Speakers
add Add new speaker

[00:00:06] Speaker 1: Panos, it's so nice to see you. Thank you so much for coming in.

[00:00:08] Speaker 2: Great to see you all.

[00:00:09] Speaker 1: I mean, you know, the wonderful thing about seeing you is that we could just spend an hour talking about music interests. And we could talk about the kind of music you listen to, the kind of music I listen to, because we're both music nuts. But I think maybe a good place to start off with, because it's where our listeners would start off, is just explaining a bit about what the Recording Academy is.

[00:00:28] Speaker 2: Most people know us as the producers of the Grammy Awards, but that is a one time a year, one day a year event. The Academy is a 365 day a year organization that is dedicated to advocacy, support, education, community building among these creatives.

[00:00:50] Speaker 1: Could you just take us inside how you think about whether or not musicians, creators are getting a fair deal from the very powerful companies who now, outside of record labels, who now dominate the industry?

[00:01:05] Speaker 2: Look, the economics of the industry have changed tremendously. And no, I mean, my view is generally that no artist should be making more from streaming services, undoubtedly. Why are they not making as much as they should? Because the economics of a streaming model are very different than the economics of an analog model of compact discs and LPs and so forth. And the way that streaming services distribute royalties and income is just very different. I won't go into all the arcane mechanics of that. And of course, as an organization, that's our foremost piece of advocacy, right? How do you get more equitable distribution of the pie? Because without creators, we have absolutely nothing. How do you actually make that happen?

[00:02:11] Speaker 1: How do you actually get a better deal for the people, as you say, who, if we don't have them creating anything, we've got no music?

[00:02:18] Speaker 2: Certainly, there is a role for legislation and arguing for better rates. And again, I don't want to get into the arcane mechanics of the way that the music copyrights work, but music is fairly unique in the sense that it has two different types of copyrights associated with a particular song. There is the sound recording, which is the actual recording that you're listening to, and that's called a mechanical right. And then there is the underlying composition, the song in itself, and the composition of that song and the composer of that song that has a different copyright, the publishing copyright.

[00:03:09] Speaker 1: Gosh, you learn a lot during this podcast. So possibly different legislation, radical legislation to sort of make it to what to sort of, I mean, you're not going to be able to get rid of that dichotomy, but to change the way copyright works.

[00:03:21] Speaker 2: It's legislation, it's advocacy, it's engagement with the technologists behind a lot of these firms. Also, we were about to talk about AI. I do believe that we do need to arrive to a point where there's transparency around what I'm consuming as a consumer. Of course.

[00:03:53] Speaker 1: Is it human or has it been manufactured by the cognitive intelligence of a machine? It seems so basic that you'd need that kitemark.

[00:04:01] Speaker 2: Yeah. I mean, look, when I go to the grocery store and I buy a jar of salsa, I can choose to get a one that's full of preservatives, or I can choose to get a homemade one. But there's something that enables me to make that decision. And that is labeling. Might that exist so that with the products that I'm consuming from a content side, music or otherwise, that enables me to make an intelligent decision.

[00:04:42] Speaker 1: I'm just going to read for our wonderful listeners, the beginning of an article on the BBC News website written by my brilliant colleagues on the 17th of January. Sienna Rose is having a good month. Three of her dusky jazz-infused soul songs are in Spotify's Viral Top 50. The most popular, a dreamy ballad called Into the Blue, has been played more than 5 million times. If she continues on this trajectory, Rose could become one of the year's hottest new stars. There's only one problem. All the signs are, all the signs indicate, sorry, she's not real. The story goes on, streaming service Deezer, which has developed tools to tackle AI music, told the BBC that, and I quote, many of her albums and songs on the platform are detected and flagged as computer generated. Look closer and you'll see the indications of an AI artist. Panos, how big a threat to the music industry is AI?

[00:05:37] Speaker 2: I think it's a massive threat. And look, there's also a big upside. We cannot downplay the threat that an artist like Sienna Rose has just described, poses to the incomes of creators. And that is at the top of everybody's mind, and certainly at the top of our mind as the recording academy. Because of the economics that I described earlier, well, the more streaming services are dominated by artists that are completely manufactured, well, it goes without saying, the less the incomes of human creators become. And there could also be a counter incentive by streaming services to necessarily identify AI music, because if you don't have to pay somebody, then great, right? How fantastic.

[00:06:41] Speaker 1: You don't even have to give Sienna royalties because there's no bank account. Is Sienna with NatWest?

[00:06:48] Speaker 2: I dread a future where 99% of the music on any of these streaming services is not human created. I don't want to experience that world. Is that world coming? I'd like to believe not. I'd like to believe that we will, as an industry, understand that that is not in the long-term benefit of everybody. So is there a world where perhaps functional music will be entirely AI driven? Maybe. But no, I don't want to experience a world where the entirety of the catalogue that I'm accessing is somehow driven by machines. That's dystopia. So what is being done?

[00:07:37] Speaker 1: What are you at the recording academy? And what is being done to fight this, as you say, tidal wave? You said a moment ago, you don't want to live in a world where 99% of what's available is this, some people call it AI slot, machine generated creation, which makes it sound godlike. And I guess in some ways, maybe that's the way we're heading. What can be done? What is being done to stop this relentless march of Sienna Rose and her ilk?

[00:08:05] Speaker 2: Well, first of all, it's all about the ethical use of other people's creations. So for me, there needs to be both legislation and a broader agreement about ethically developed LLMs or large language models. Then I do think that the downright stealing of identity, stealing of intellectual property, and the creation of content that is purely meant to mislead, there should be some legislation that makes that illegal. Undoubtedly. I cannot be listening to a song that I think is made by you name the artist, Drake, Bad Bunny, Rosalia, and think that it's her. That's not okay. Any more that it's not okay that I go to a Gordon Ramsay restaurant that's not a Gordon Ramsay restaurant. That's sort of trading under false pretenses. To me, that is just baseline. Beyond that, legislation that bans the unauthorized use of my image and likeness and voice that enables a more ethically created use of intellectual property without stalling or thwarting technology or progress, nothing will do that. All that's important. So as an organization, we're, again, right now we're based out of the United States, but we spend a lot of time educating legislators and on Capitol Hill to advance legislation that fundamentally protects human creators. Now, I also, I'm a big believer in agency, in giving agency to the creator to choose what happens to that. Look, if I want to outsource my voice and give millions of young creators the ability to create wholly new expressions with my voice, that's the choice that I'm making. Go at it. If you choose that you want to get compensated every time, I'd like to see the rise of systems and the ability for me as a creator that's outsourced my voice and all these other creators of a different ilk than me are making new products. Great. If there's a way for me to be remunerated with that, fantastic. Fantastic. So I see it as maybe I'm beginning to go down the path where I do see opportunities here for actually further income growth. And you've seen the use of such open source models in all kinds of industries that have catalyzed innovation and have grown collective incomes. But I still think we first have to cross a giant chasm. And if we don't do it right, then I do think that it can be an existential threat. And as a recording academy, we're not going to allow that.

ai AI Insights
Arow Summary
Interview excerpt with Panos from the Recording Academy explaining the Academy’s year-round role beyond the Grammys: advocacy, support, education, and community for music creators. Discussion centers on creators’ compensation in the streaming era, how royalty economics differ from physical media, and the dual-copyright structure in music (sound recording/mechanical rights and composition/publishing). Panos argues creators should earn more, with solutions including legislation, better rates, and engagement with technology companies. The conversation then shifts to AI-generated music as a major threat—citing an apparent AI artist gaining millions of streams—and warns that AI content could depress human creators’ income and incentivize platforms to avoid detection. Proposed responses include transparency/labeling for consumers, laws against deceptive impersonation and unauthorized use of identity/likeness/voice, ethical development of models, and creator “agency” to opt in/out and be compensated, while acknowledging AI could also create new revenue opportunities if governed properly.
Arow Title
Recording Academy on Streaming Royalties and AI Music Risks
Arow Keywords
Recording Academy Remove
Grammys Remove
music industry Remove
streaming royalties Remove
artist compensation Remove
copyright Remove
mechanical rights Remove
publishing rights Remove
legislation Remove
advocacy Remove
AI-generated music Remove
deepfakes Remove
transparency Remove
labeling Remove
creator agency Remove
ethical AI Remove
platform incentives Remove
Arow Key Takeaways
  • The Recording Academy operates year-round, focusing on advocacy and support for music creators beyond the Grammy Awards.
  • Streaming’s royalty economics differ significantly from physical media, contributing to lower artist earnings.
  • Music has two core copyrights per song: sound recording (mechanical) and composition (publishing), complicating policy and payouts.
  • Improving creator compensation likely requires legislation, rate-setting reforms, and engagement with tech and platforms.
  • AI-generated music poses a significant income threat by increasing non-human catalog share and potentially reducing payouts to humans.
  • Platforms may have weak incentives to police AI content if it lowers costs, making transparency and enforcement crucial.
  • Consumers may need clear labeling to distinguish human-made from AI-generated content.
  • Legislation should target deceptive impersonation and unauthorized use of voice/likeness while allowing ethical innovation.
  • Creator agency and opt-in licensing/compensation systems could turn some AI uses into new revenue streams.
  • Poor governance of AI could be existential for human creators; proactive action is needed.
Arow Sentiments
Neutral: Measured, policy-focused tone balancing concern about inequitable streaming economics and AI-related threats with cautious optimism about potential upsides if regulation, transparency, and compensation mechanisms are implemented.
Arow Enter your query
{{ secondsToHumanTime(time) }}
Back
Forward
{{ Math.round(speed * 100) / 100 }}x
{{ secondsToHumanTime(duration) }}
close
New speaker
Add speaker
close
Edit speaker
Save changes
close
Share Transcript