WEBVTT - SoundExchange President and CEP Michael Huppe Talks AI in Music Industry

0:00:00.240 --> 0:00:03.000
<v Speaker 1>Carol. You'll recall just last week at the Bloomberg Tech

0:00:03.080 --> 0:00:04.400
<v Speaker 1>Summit we sat down with hit Boy.

0:00:04.640 --> 0:00:05.720
<v Speaker 2>Was it one we could go today?

0:00:05.760 --> 0:00:07.760
<v Speaker 1>It was one we could go today. Three time Grammy

0:00:07.800 --> 0:00:09.920
<v Speaker 1>Award at winning artist producer is known for shaping the

0:00:09.920 --> 0:00:13.080
<v Speaker 1>sounds of jay Z, Beyonce, Nosdrich and more. Check out

0:00:13.119 --> 0:00:15.880
<v Speaker 1>his instagram for some fun shots from last week behind

0:00:15.880 --> 0:00:16.320
<v Speaker 1>the scenes.

0:00:16.960 --> 0:00:17.800
<v Speaker 2>It's a lot of fun.

0:00:18.440 --> 0:00:20.119
<v Speaker 1>The topic of our chat was all about AI and

0:00:20.200 --> 0:00:22.720
<v Speaker 1>using AI to create music. He's actually game. He says

0:00:22.720 --> 0:00:25.640
<v Speaker 1>that it's music the same way he was using fruity

0:00:25.680 --> 0:00:27.880
<v Speaker 1>loops earlier in his career. It's just you got to

0:00:27.920 --> 0:00:30.760
<v Speaker 1>adjust with the times. That is not, in fact, the

0:00:30.800 --> 0:00:34.320
<v Speaker 1>AI problem with music that Michael Huppy, the Presidency of

0:00:34.360 --> 0:00:38.199
<v Speaker 1>Sound Exchange, has identified. Sound Exchange is this nonprofit. It's

0:00:38.200 --> 0:00:42.440
<v Speaker 1>owned by the radio industry. It calls itself the music

0:00:42.479 --> 0:00:45.640
<v Speaker 1>industry excuse me, calls itself the largest global neighboring rights

0:00:45.720 --> 0:00:47.960
<v Speaker 1>organization in the world. It says it's collected and distributed

0:00:48.200 --> 0:00:51.279
<v Speaker 1>more than thirteen billion dollars in digital performance royalties to date,

0:00:51.320 --> 0:00:53.440
<v Speaker 1>on behalf of more than eight hundred thousand music creators.

0:00:53.479 --> 0:00:56.560
<v Speaker 1>Michaelhuppy joins us here in the Bloomberg Interactive Broker's studio.

0:00:56.880 --> 0:00:59.400
<v Speaker 1>It has been quite a bit since we've talked to you.

0:01:00.000 --> 0:01:01.600
<v Speaker 1>I want to get to the AI problem that you've

0:01:01.600 --> 0:01:04.760
<v Speaker 1>identified before. We do that, though Sound Exchange not a

0:01:04.760 --> 0:01:07.240
<v Speaker 1>household name for a lot of people, even though every

0:01:07.319 --> 0:01:11.000
<v Speaker 1>day they're interacting with the product that it touches. Where

0:01:11.040 --> 0:01:13.600
<v Speaker 1>do you sit? Remind everybody where you sit in sort

0:01:13.640 --> 0:01:15.800
<v Speaker 1>of the artists get paid when you listen to the

0:01:15.880 --> 0:01:18.320
<v Speaker 1>radios or when you listen to certain radio.

0:01:18.360 --> 0:01:20.480
<v Speaker 3>Should sure, first off, thanks for having me back. It

0:01:20.560 --> 0:01:22.600
<v Speaker 3>was always great to be here. Sound Exchange is a

0:01:22.640 --> 0:01:25.560
<v Speaker 3>company that represents the entire recorder music industry, all artists,

0:01:25.600 --> 0:01:28.080
<v Speaker 3>all record labels. We sit between a lot of the

0:01:28.120 --> 0:01:33.039
<v Speaker 3>streaming radio services think Sirius, XM, Pandora, iHeart Online.

0:01:32.600 --> 0:01:34.840
<v Speaker 1>And not Spotify Spotify, but a.

0:01:34.840 --> 0:01:36.760
<v Speaker 3>Lot of the ones that are not interactive radio. And

0:01:36.800 --> 0:01:39.039
<v Speaker 3>then on the other side you have all the labels,

0:01:39.080 --> 0:01:42.680
<v Speaker 3>all the artists, and we help pay out all of

0:01:42.720 --> 0:01:45.440
<v Speaker 3>the all of the royalties basically that are created by

0:01:45.480 --> 0:01:47.080
<v Speaker 3>all these services to the whole record and.

0:01:47.200 --> 0:01:50.920
<v Speaker 1>So every time somebody listens to Taylor Swift on Pandora,

0:01:51.000 --> 0:01:55.040
<v Speaker 1>for example, you then handle writing the check to.

0:01:55.600 --> 0:01:59.280
<v Speaker 3>Who exactly, to Taylor Swift or her management company or

0:01:59.280 --> 0:02:01.240
<v Speaker 3>whoever may be. We also send half that money to

0:02:01.280 --> 0:02:06.120
<v Speaker 3>a record label. Okay, So it's a centralized place, makes

0:02:06.160 --> 0:02:09.400
<v Speaker 3>it very efficient. You know, thousands of services, over eight

0:02:09.520 --> 0:02:12.680
<v Speaker 3>hundred thousand accounts we have, and it's one efficient, centralized

0:02:12.720 --> 0:02:14.480
<v Speaker 3>way to get all of that money out to the

0:02:14.520 --> 0:02:17.480
<v Speaker 3>artists and labels who form the basis of these products.

0:02:17.560 --> 0:02:19.600
<v Speaker 2>I'm curious, are you paying out more than you were

0:02:20.120 --> 0:02:20.639
<v Speaker 2>last year?

0:02:20.680 --> 0:02:20.760
<v Speaker 1>Like?

0:02:20.880 --> 0:02:24.000
<v Speaker 2>Is it growing or is there more competition? It's an other.

0:02:24.200 --> 0:02:28.640
<v Speaker 3>Yeah, it's pretty it's consistently grown. I would say fifteen

0:02:28.680 --> 0:02:31.080
<v Speaker 3>twenty years ago we were paying out twenty five million.

0:02:31.160 --> 0:02:33.320
<v Speaker 3>Now we pay north of a billion dollars a year.

0:02:33.360 --> 0:02:37.280
<v Speaker 3>It's been pretty consistent growth due to a lot of reasons. Obviously,

0:02:37.639 --> 0:02:40.320
<v Speaker 3>streaming is up and the way people consume music, you know,

0:02:40.360 --> 0:02:43.160
<v Speaker 3>it's all collapsing onto the phone. So it's been a

0:02:43.160 --> 0:02:44.520
<v Speaker 3>pretty banner of fifteen years.

0:02:44.560 --> 0:02:48.320
<v Speaker 2>What do people get paid like for an album a song?

0:02:48.360 --> 0:02:51.720
<v Speaker 2>I'm always curious or does it vary? Like how is it?

0:02:51.800 --> 0:02:54.160
<v Speaker 3>Well? It varies and it depends on the product, right,

0:02:54.280 --> 0:02:57.920
<v Speaker 3>So satellite radio, for instance, a subscription based business. They

0:02:58.120 --> 0:03:00.720
<v Speaker 3>we get a percentage of revenue based on you know,

0:03:00.800 --> 0:03:04.400
<v Speaker 3>relevant revenue. But when you're talking about webcasting, for instance,

0:03:04.440 --> 0:03:08.920
<v Speaker 3>typically every stream to every set of years is a

0:03:09.000 --> 0:03:12.720
<v Speaker 3>fraction of a penny and it's all set every five years.

0:03:12.760 --> 0:03:14.400
<v Speaker 3>And you know, that doesn't sound like a lot of money,

0:03:14.400 --> 0:03:15.880
<v Speaker 3>but as I said, over the course of a year,

0:03:15.880 --> 0:03:19.000
<v Speaker 3>it adds up to over a billion dollars. We are alone,

0:03:19.040 --> 0:03:21.639
<v Speaker 3>you know, twelve or twelve ish fifty percent or so

0:03:21.800 --> 0:03:23.680
<v Speaker 3>of the whole US recorded music revenue.

0:03:23.720 --> 0:03:26.200
<v Speaker 1>I know it's past his bedtime, but Tom Keene, he

0:03:26.560 --> 0:03:26.880
<v Speaker 1>the host of.

0:03:26.840 --> 0:03:29.560
<v Speaker 2>Bloomberg, thinks he has you.

0:03:29.600 --> 0:03:31.760
<v Speaker 1>Know, he has a new England folk album out, I know,

0:03:31.919 --> 0:03:34.760
<v Speaker 1>nineteen ninety two searching for Ward in June, So you

0:03:34.800 --> 0:03:37.360
<v Speaker 1>could ask him about the royalties.

0:03:36.920 --> 0:03:39.280
<v Speaker 3>He should register for sound Exchange because if he gets

0:03:39.280 --> 0:03:41.360
<v Speaker 3>play on any of those platforms, we have money for him.

0:03:41.400 --> 0:03:41.880
<v Speaker 1>I love it.

0:03:42.000 --> 0:03:43.440
<v Speaker 2>So where does AI fit into.

0:03:43.320 --> 0:03:45.680
<v Speaker 3>All of this? AI is a fascinating part of the

0:03:45.680 --> 0:03:48.320
<v Speaker 3>industry coming up. It's something that has a lot of

0:03:48.440 --> 0:03:51.240
<v Speaker 3>danger but also a lot of potential. So you know,

0:03:51.680 --> 0:03:54.800
<v Speaker 3>I think we should all view AI as something that

0:03:54.920 --> 0:03:58.840
<v Speaker 3>can make really potential revenue growth for the industry, new

0:03:58.840 --> 0:04:00.960
<v Speaker 3>products for consumers, but it has to be rolled out

0:04:00.960 --> 0:04:03.880
<v Speaker 3>the right way. Specifically, we need to make sure that

0:04:04.000 --> 0:04:06.360
<v Speaker 3>human creators are protected. There need to be guardrails so

0:04:06.400 --> 0:04:08.800
<v Speaker 3>that doesn't steamroll over the whole creative industry.

0:04:08.840 --> 0:04:11.640
<v Speaker 1>Well, what's the problem that right now artists are facing

0:04:12.000 --> 0:04:13.760
<v Speaker 1>when it comes to AI generated music.

0:04:13.960 --> 0:04:17.040
<v Speaker 3>I mean, there's a lot of potential problems. One is,

0:04:17.080 --> 0:04:19.600
<v Speaker 3>you know, when you think about the streaming world, there

0:04:19.640 --> 0:04:23.200
<v Speaker 3>are services now that are reporting seventy five thousand uploads

0:04:23.360 --> 0:04:27.560
<v Speaker 3>a day a day of new recordings, eighty percent or

0:04:27.560 --> 0:04:29.719
<v Speaker 3>plus of them are AI music. When you dump what.

0:04:29.839 --> 0:04:32.560
<v Speaker 1>Does that make it onto Pandora? Does it make it

0:04:32.560 --> 0:04:33.520
<v Speaker 1>onto serious excent?

0:04:33.880 --> 0:04:36.600
<v Speaker 3>It is making it into some of these streaming services,

0:04:36.640 --> 0:04:39.120
<v Speaker 3>maybe not at the scale that you would expect yet,

0:04:39.120 --> 0:04:40.680
<v Speaker 3>but I mean there have been reports of you know,

0:04:40.680 --> 0:04:43.039
<v Speaker 3>whether it's one to two or three percent. Sometimes people

0:04:43.160 --> 0:04:45.640
<v Speaker 3>estimate that it's up to ten percent. There's also the

0:04:45.680 --> 0:04:48.720
<v Speaker 3>ability to streaming fraud is another big issue. There are

0:04:48.800 --> 0:04:52.599
<v Speaker 3>cases where people use AI tracks, set up bots around

0:04:52.640 --> 0:04:58.120
<v Speaker 3>and they siphon away payment from the pipeline that would

0:04:58.120 --> 0:05:00.680
<v Speaker 3>otherwise go to real artists and real record label. It's

0:05:00.680 --> 0:05:03.520
<v Speaker 3>fraud basically straight up fraud. There are people in prosecuted.

0:05:03.520 --> 0:05:05.240
<v Speaker 3>There's a famous case in the Southern District of New

0:05:05.320 --> 0:05:06.600
<v Speaker 3>York that was prosecuted recently.

0:05:06.640 --> 0:05:09.480
<v Speaker 1>What's a bigger issue for AI generated music is that

0:05:09.560 --> 0:05:12.520
<v Speaker 1>the fraud or is it the uploading of AI generated music?

0:05:12.640 --> 0:05:12.800
<v Speaker 2>Right?

0:05:12.839 --> 0:05:15.320
<v Speaker 3>I mean it's both the uploading of I mean fraud

0:05:15.440 --> 0:05:18.280
<v Speaker 3>is something that I think we're tackling better as an industry.

0:05:18.279 --> 0:05:20.640
<v Speaker 3>There are company is dedicated to it now. I think

0:05:20.680 --> 0:05:22.000
<v Speaker 3>if you were to talk to a lot of the

0:05:22.080 --> 0:05:24.160
<v Speaker 3>artists or the labels, they want to make sure that

0:05:24.640 --> 0:05:28.040
<v Speaker 3>when as AI expands that it doesn't squeeze out the

0:05:28.120 --> 0:05:32.640
<v Speaker 3>human element, you know, human creativity, human input is critical.

0:05:32.839 --> 0:05:37.280
<v Speaker 2>But will you get a percentage of AI created songs?

0:05:37.400 --> 0:05:39.880
<v Speaker 2>Like it's like, how is that?

0:05:39.920 --> 0:05:40.520
<v Speaker 1>What this like?

0:05:40.640 --> 0:05:42.239
<v Speaker 2>I'm just curious how you guys are jockey?

0:05:42.440 --> 0:05:47.200
<v Speaker 3>Well, so so sound exchange AI created songs. Actually, this

0:05:47.279 --> 0:05:50.159
<v Speaker 3>is going to be interesting. They actually don't merit a

0:05:50.160 --> 0:05:52.920
<v Speaker 3>payment because they're not protectable under the law. And under

0:05:53.040 --> 0:05:56.640
<v Speaker 3>US law, something is holy AI created. It's literally not copyrightable.

0:05:57.080 --> 0:06:00.400
<v Speaker 3>So this isn't about more money coming through sound Change

0:06:00.440 --> 0:06:03.680
<v Speaker 3>or not. It's basically about looking to protect the creative community,

0:06:03.720 --> 0:06:06.520
<v Speaker 3>protect creators and make sure that these companies, I mean

0:06:06.520 --> 0:06:09.800
<v Speaker 3>AI companies are made are going to will make trillions

0:06:09.800 --> 0:06:13.479
<v Speaker 3>of dollars and they part of that. Part of that

0:06:13.560 --> 0:06:16.440
<v Speaker 3>business models built on the backs of artists and labels

0:06:16.480 --> 0:06:17.640
<v Speaker 3>when they scraped their content.

0:06:17.880 --> 0:06:20.160
<v Speaker 1>So but if it's not good, nobody's going to listen

0:06:20.160 --> 0:06:20.440
<v Speaker 1>to it.

0:06:21.360 --> 0:06:23.360
<v Speaker 3>Well, do you think a music is good?

0:06:23.440 --> 0:06:25.920
<v Speaker 1>I've never known till I don't know. I've never listened

0:06:25.960 --> 0:06:26.120
<v Speaker 1>to it.

0:06:26.640 --> 0:06:28.560
<v Speaker 2>Sometimes did it like a study, like pick the AI

0:06:28.720 --> 0:06:30.760
<v Speaker 2>generated song, and I was pretty impressed at how good

0:06:30.760 --> 0:06:32.680
<v Speaker 2>it was. But I mean, if it if it sounds

0:06:32.760 --> 0:06:36.800
<v Speaker 2>like vocals, I think it did have vocals.

0:06:36.920 --> 0:06:39.920
<v Speaker 3>Yeah. AI music has come a long way. Say, when

0:06:39.960 --> 0:06:41.880
<v Speaker 3>it first started out, you could pretty much tell. But

0:06:42.560 --> 0:06:45.200
<v Speaker 3>the stuff that AI is putting out now is pretty impressive.

0:06:45.240 --> 0:06:47.480
<v Speaker 3>But but it's great if they do that. I want,

0:06:47.520 --> 0:06:50.479
<v Speaker 3>I want to be clear. You know, AI has this risk,

0:06:50.520 --> 0:06:52.640
<v Speaker 3>but it also has a lot of potential to grow

0:06:52.720 --> 0:06:56.039
<v Speaker 3>the industry, grow revenue. And one thing about what's happening

0:06:56.080 --> 0:06:58.520
<v Speaker 3>now compared to say, the napster days, I mean you

0:06:58.520 --> 0:07:00.880
<v Speaker 3>you guys remember the napster days. It's very different now.

0:07:01.000 --> 0:07:03.479
<v Speaker 3>Is the recording industry is leaning in. They're working with

0:07:03.560 --> 0:07:06.960
<v Speaker 3>the AI companies. Their AI companies are licensing, and they're

0:07:06.960 --> 0:07:09.279
<v Speaker 3>trying to work together to develop these models.

0:07:09.279 --> 0:07:11.600
<v Speaker 2>I mean computers. Technology has been part of music for

0:07:11.640 --> 0:07:14.120
<v Speaker 2>a long time. What do they call those things when

0:07:14.160 --> 0:07:15.080
<v Speaker 2>people can't sing and.

0:07:15.040 --> 0:07:16.760
<v Speaker 1>They auto tune auto ten.

0:07:17.200 --> 0:07:20.080
<v Speaker 2>There you go, Michael Happy, Thank you so much, President,

0:07:20.280 --> 0:07:21.800
<v Speaker 2>CEO of Sound Exchange,