Field Dispatch
Matt Connarton Unleashed: Celine Dion vs. AI
Speaker 1: We've been talking a lot about AI on the show.
Speaker 2: Last week, we did we did our experiment with Souno
Speaker 2: dot com where we made some tracks, and apparently it
Speaker 2: is even easier than I could have imagined.
Speaker 3: Too easy, stupid easy. Yeah, like really really too easy.
Speaker 3: If you didn't hear the show last week, seriously, go
Speaker 3: back and listen to it. Yeah, we didn't give it
Speaker 3: a ton, just some basic info and it actually spit
Speaker 3: out some decent music. And it's concerning because how do
Speaker 3: you know that's AI? Right, Like, you need AI to
Speaker 3: find out that it's AI, and then what music did
Speaker 3: they use to learn that? You're right? Yeah, nothing coming
Speaker 3: out of AI. Didn't have a human On the other end,
Speaker 3: they're they're taking all of the art, all of the
Speaker 3: creations from music genres across the spectrum, from artists everywhere,
Speaker 3: throwing that in a computer and having it spit It's
Speaker 3: to me, it's still using other people's work. So that's
Speaker 3: my my. There has to be a way to tell
Speaker 3: the difference between real human creativity and AI.
Speaker 2: Yeah, if you go back and listen to last week's show,
Speaker 2: you can check out some of the tracks that we created,
Speaker 2: just with a simple prompt and what it was able
Speaker 2: to spit out. Suno dot Com is a service and
Speaker 2: they're actually they're free. The free version of the service
Speaker 2: is pretty generous. They give you, you know, you can make
Speaker 2: ten songs a month, which might not sound like a lot,
Speaker 2: but some of these services, you know, they they try
Speaker 2: to get you into a free trial and get you
Speaker 2: hooked on it so then you'll be willing to pay
Speaker 2: for it every month.
Speaker 1: But what Suno does is.
Speaker 2: The free option that they offer is actually pretty good.
Speaker 2: But if you want to hear, so you can hear.
Speaker 2: What we did live on the air was something we'd
Speaker 2: wanted to get to for a while, just kind of
Speaker 2: an experiment of creating tracks using one of these AI services,
Speaker 2: and we created several and and of course I did
Speaker 2: put the the mc hammer distracts because of course my
Speaker 2: long running feud with mc hammer going back six years now,
Speaker 2: long time listeners, we'll know the six years now. Yes
Speaker 2: we did, uh, I did the po. We ended up
Speaker 2: making four distracts using Souno dot com and I did
Speaker 2: post them all on YouTube as well, so in your
Speaker 2: face Hammer, it's yeah, that's right.
Speaker 3: So people did an article recently about Celine Dion uh huh,
Speaker 3: So the AI is crossing all gendres as far as
Speaker 3: what it's doing. But apparently Selene Dion did an Instagram
Speaker 3: post this past Friday. Just yesterday calling out AI and
Speaker 3: saying that her voice has been allegedly used in an
Speaker 3: unsanctioned way that she did not approve of. That it
Speaker 3: sounds like it's her giving a performance, But it's not
Speaker 3: pretty upsetting when you can consider the fact that, I mean,
Speaker 3: for one thing, she's an amazing artist, but this woman
Speaker 3: is going through enough right now without this on top
Speaker 3: of it all. I mean, look, can we just say that,
Speaker 3: because I'm really this upset me to see, of all people,
Speaker 3: you've got to go after somebody who's literally fighting for
Speaker 3: their life. That's pretty down and low and disgusting if
Speaker 3: you ask me, but to quota it said, It has
Speaker 3: come to our attention that unsanctioned AI generated music purporting
Speaker 3: to contain Celine Dion's musical performances and names and likenesses
Speaker 3: excuse me and name and Likeness is currently circulating online
Speaker 3: and across various digital service providers. Please be advised that
Speaker 3: these recordings are fake and not approved, and are not
Speaker 3: songs from her official discography.
Speaker 2: So we should clarify this because I think I understand
Speaker 2: you know, I only had a chance to glance at
Speaker 2: this before, but I so because my first thought was, well,
Speaker 2: how do you prove that her copyrighted work has been
Speaker 2: used to train these AI models? And then I realized
Speaker 2: that's not exactly what's happening here, right, people are creating
Speaker 2: AI versions and claiming it is her.
Speaker 1: Yes, okay, yeah, this.
Speaker 3: Is actually they're putting stuff out saying it's her, yeah,
Speaker 3: and it's it's not by any you know, so this
Speaker 3: so she just got blindsided by this.
Speaker 2: Well you know what this reminds me of. And this
Speaker 2: was over a year ago, I think now that we
Speaker 2: talked about it, somebody created on YouTube they posted an
Speaker 2: AI version of a brand new George Carlin album. Yes,
Speaker 2: remember we talked about that, and and that was interesting.
Speaker 2: We actually played a little bit of it. We played
Speaker 2: the very beginning of it on the show because I'd
Speaker 2: already listened to it, so I knew where the first
Speaker 2: swear was, so I knew how far, how far into
Speaker 2: it I could play because I wanted to give people
Speaker 2: an idea of what it sounded like and very real sound.
Speaker 2: It did sound very real, and not only that, I
Speaker 2: ended up listening to the entire hour. It actually was
Speaker 2: funny too, Like it wasn't like, okay, it sounded like
Speaker 2: his voice, but the jokes were really lame. Like I
Speaker 2: actually laughed listening to that special Which is part of
Speaker 2: why some people at the time were a little bit
Speaker 2: skeptical and probably still are that because the people who
Speaker 2: created it claimed that it was entirely AI generated, but
Speaker 2: some people listened to it and said, hmm, yeah, the
Speaker 2: voice maybe, but some of those jokes, there's no way
Speaker 2: those were written by AI. Those were inserted by who
Speaker 2: ever made it, but who knows. But yeah, you can make.
Speaker 3: Very realistic That's something that anybody can really say.
Speaker 1: Yeah, but you can make very realistic work with it.
Speaker 3: That's smart of what you give it. So if you
Speaker 3: fed everything to it of George Carlin, I actually don't
Speaker 3: doubt that it could write George Carlin jokes because it
Speaker 3: would have learned from that. You know, it's a computer.
Speaker 3: What you feed it, Yeah, right, all the bigga ones
Speaker 3: and zeros.
Speaker 2: But humor is Humor is different though. That's why I'm
Speaker 2: also I'm one of the people who is skeptical but
Speaker 2: who knows.
Speaker 1: We'll probably never know the truth.
Speaker 3: I got you, Yeah, but I mean in April of
Speaker 3: last year, there were over two hundred artists that got
Speaker 3: together and created an Artist Writes Alliance, and they sign letters.
Speaker 3: And we're trying to get people to pay attention in
Speaker 3: saying that you know that they want they want laws
Speaker 3: to protect them, and they want AI developers to have
Speaker 3: some freaking morals, I think, and not steal from them,
Speaker 3: because it really is. It's stealing somebody else's are you're
Speaker 3: stealing theon Celene Dion's voice. That's so wrong. It's more
Speaker 3: wrong by because of who they chose to pick on
Speaker 3: as somebody who's literally fighting for their If you don't know,
Speaker 3: Selene Dion has what's called stiff person syndrome. That's why
Speaker 3: she can't perform anymore. It only does so on very
Speaker 3: very rare occasions. It's extraordinarily painful. She suffers immense, just
Speaker 3: chronic every day suffering that you cannot even imagine. And
Speaker 3: that's who they pick. They steal her voice, put music
Speaker 3: out there, claim it's her trying to make money off
Speaker 3: of her.
Speaker 2: Well, also want to make a clear distinction in here
Speaker 2: again going back to what's really happening in this instance,
Speaker 2: because you know, we've had a lot of conversations about uh,
Speaker 2: generative AI, these large learning models. You know, they suck
Speaker 2: up all this information and all this art, all this
Speaker 2: creativity and then spit out what they spit out. And
Speaker 2: we've we've had conversations debating about whether that's okay or not.
Speaker 2: But this is different in that this because this is
Speaker 2: not there's no there's no ambiguity here, because this is
Speaker 2: fraud if you're creating something, because you can take the
Speaker 2: AI completely out of it. Let's let's say it's just
Speaker 2: somebody who does a really good impression of Selene Deon
Speaker 2: recording a song and then claiming that it is Selene Deon.
Speaker 1: That's out that's outright fraud.
Speaker 2: If you create something and then and then and you
Speaker 2: claim it's something that it's not, that's that's fraud. That
Speaker 2: is there's no ambiguity there.
Speaker 3: But the only differences in this instance, because it's AI,
Speaker 3: they can make it sound just like her. It's not
Speaker 3: just kind of like it's just like you know, because
Speaker 3: computers can imitate everything, right, and these and what the
Speaker 3: Artist Rights Alliance is trying to do is protect each other.
Speaker 3: So when I'm talking about over two hundred, like really
Speaker 3: big names signed onto this, like Billie Irish and and
Speaker 3: She's imagine dragons, Sheryl Crow, Stevie Wonder, a whole lot
Speaker 3: of people, and they they're from their perspective. And I
Speaker 3: want to give this quote because I think it's a
Speaker 3: great quote. They said quote, Unfortunately, some platforms and developers
Speaker 3: are employing AI to sabotage creativity and undermined artists, songwriters,
Speaker 3: musicians and rights holders. And I think that that's a
Speaker 3: very very perfect statement. It is sabotaging their creativity. It's
Speaker 3: theft and fraud to steal somebody else's labors turn it
Speaker 3: into whatever you want to turn it into. You know
Speaker 3: that it's and you know what, Artists have enough problems
Speaker 3: these days making money off their own music. I mean
Speaker 3: talk about that before. What they make less than what
Speaker 3: is it.
Speaker 1: Whatever Spotify off right, Yeah, it's not even.
Speaker 3: In the old days, you printed an album you made
Speaker 3: your money off of in part apples sales, album sales,
Speaker 3: and then touring. These days you're making more of your
Speaker 3: money off of touring because you can't make as much
Speaker 3: money off of album sales.
Speaker 2: Although although what a lot of people don't realize is
Speaker 2: it's kind of it's it's it's uh. In terms of
Speaker 2: trying to earn revenue from streaming, it's kind of a
Speaker 2: worse version of what was already the model, in the
Speaker 2: sense that you never really made that.
Speaker 1: Much off of album sales. This is something people don't know.
Speaker 2: And I've given this example before, but I remember when
Speaker 2: I was a kid when I first started to realize,
Speaker 2: because I was fascinated by all you know, all this stuff.
Speaker 2: When I mean AI, we didn't have yet. But when
Speaker 2: I was a kid, I was always very interested in
Speaker 2: the music industry and how it works. And I remember
Speaker 2: reading an interview with meat Loaf and he was talking
Speaker 2: about Back Out of Hell and of course, which by
Speaker 2: now is probably sold and it might be it may
Speaker 2: have surpassed one hundred million at this point, but he
Speaker 2: was talking about how he didn't really make that much money,
Speaker 2: like he said, you know, he got like a quarter
Speaker 2: a quarter, like twenty five cents a quarter per unit sold.
Speaker 1: Wow, you know, And but that's that's not unusual.
Speaker 3: Although music didn't get in music how to be transferred
Speaker 3: from one person or another in an album. Yeah, for
Speaker 3: at a cassette. You couldn't just electronically, but it.
Speaker 2: Wasn't but it wasn't as much. That's why I like
Speaker 2: sometimes every once in a while you'll hear very successful.
Speaker 2: I remember Kurt Cobain, you know, uh, of course he
Speaker 2: took himself away from us very young. But I remember
Speaker 2: when Nirvana first got really big. He made he made
Speaker 2: a comment in an interview about you know, people think
Speaker 2: because we're famous, we're rich, and it's it doesn't work
Speaker 2: that way. He said, you'd be surprised how broke we are.
Speaker 2: And TLC we've talked about that a great example. They
Speaker 2: were triple platinum and broke. The money has always been
Speaker 2: in touring, the big money, and now even touring is
Speaker 2: is as that has become more cost prohibitive. Now the
Speaker 2: money really is in merch and and of course licensing.
Speaker 2: You know, if you license your music, say to a
Speaker 2: for a film or a TV series or something, then
Speaker 2: you can make a lot of money. There's always been
Speaker 2: a lot of money in licensing and that persists.
Speaker 3: But whatny do you think they make in licensing, Like
Speaker 3: when you're a venue having a license to be able
Speaker 3: to have an artist come in and play certain songs
Speaker 3: venues can't have can only have certain music that they're
Speaker 3: license for. Do they get the artists make money off
Speaker 3: of that or are you really something that Well, yeah,
Speaker 3: it doesn't get them much.
Speaker 1: It doesn't get them much.
Speaker 2: I mean those performing rights organizations like as CAP and BMI. Yeah,
Speaker 2: they're supposed to make sure that you get paid. But
Speaker 2: but but for that it's it's yeah, we're not we're
Speaker 2: talking about yeah, Spotify money.
Speaker 3: The bottom line, it's really hard for artists to be
Speaker 3: able to make money. So they have AI ripping off
Speaker 3: artists voices or you know, creating songs that they didn't
Speaker 3: even sing. Yeah, this is this is not good and
Speaker 3: I I mean, I do believe and I do agree
Speaker 3: with that. I do believe that it does. They are
Speaker 3: correct in saying it's sabotage and creativity. It's stealing creativity.
Speaker 3: It's stealing the life's blood of an artist. You know,
Speaker 3: whatever I put on a piece of canvas came out
Speaker 3: of my me, right.
Speaker 1: I wouldn't you know what I mean?
Speaker 3: You think about when I think about it that way,
Speaker 3: it's like a musician's song is their baby, that's their soul.
Speaker 3: You know what it actually brings up is. I know
Speaker 3: it's slightly off, but when with Wren the artist, Wren
Speaker 3: the rapper Rent he used a piece of music, he
Speaker 3: spent like tune of dollars to buy it, and Kujo
Speaker 3: another rapper or something had ownership over that piece, and
Speaker 3: when he saw Wren made so much money, he went
Speaker 3: and tried to sue him for it.
Speaker 2: Yeah, what a Yeah, we never really talked about it
Speaker 2: on the show, but yeah, that turned into quite a mess.
Speaker 3: Oh but the But if you want to hear a
Speaker 3: really great disk track, you got to listen to Wren's
Speaker 3: disk track on Kuja.
Speaker 2: But that wasn't even even an AI situation. It was
Speaker 2: over a sample was created.
Speaker 3: Yeah, this was trying and in in the disk track
Speaker 3: and what he would say and what Ren would say
Speaker 3: off airs. He's trying to steal my creativity. He's trying
Speaker 3: to steal my creat That song was so personable and
Speaker 3: so much of his own soul and life blood in there.
Speaker 3: To have somebody come along and say that's not your creativity,
Speaker 3: it's mine. From this two second whatever that he bought
Speaker 3: and paid for to use this it's just a sound bite,
Speaker 3: it's just a clip, not the whole song. It didn't,
Speaker 3: you know, it's nothing like that. It's just a particular
Speaker 3: looping sound. Yeah, And he bought and paid for it correctly.
Speaker 3: And this guy, you know, I want money, right, I
Speaker 3: see you making money. I want that money. Trying but
Speaker 3: he Wren said the same thing. You're trying to steal
Speaker 3: my creativity, my creation, and that it so if you
Speaker 3: look at it in all these different situations, it's it is.
Speaker 3: It's stealing the very soul, lifeblood of an artist.
Speaker 1: Yeah,
Podbean