Half of New Songs are AI, but SO ARE THE LISTENERS ($8 million dollar loophole)

On a recent episode, we covered how Anthropic recently had to settle a $1.5 billion civil suit for illegally downloading millions of books that they then fed into their AI model for training purposes. That, to date, is the single largest copyright infringement settlement in American history. This past weekend, however, over 30 music companies filed a separate class action lawsuit alleging Anthropic illegally downloaded song lyrics from those same torrent websites and intentionally removed copyright information to again feed it into the system and train the AI model. How this particular new case shakes out could very well dictate whether the live music industry is still a thing a generation from now, or if we're gonna have to just drown in a sea of AI musical sloth. And so let's dive into it right after you smash those like and subscribe buttons so the video can be picked up by the YouTube algorithm and shared out to ever more people. Thank you so much for that. Now diving into it, the Anthropic case was filed on August the 28th by Sony Music Publishing as well as Warner Chapel Music alongside over 30 different music production companies and they collectively filed a civil suit in California against Anthropic and their leadership team. The complaint alleges that Anthropic breached American copyright law by torrenting, scraping and downloading lyrics from tens of thousands of copyrighted songs to train its family of clawed AI models. Publishers in this case are seeking damages that could run into the billions of dollars as they're seeking reimbursement of up to $150,000 per song. Now we spoke with Mr. Kirk Sigmund, who is the chief innovation attorney and the founding partner at Keldon, which is a DC based intellectual property, AI and technology law firm. And we asked him how this new music lawsuit overlaps with a previous settlement where Anthropic was accused of illegally downloading over 7 million books, some of which included copywritten lyrics. There seems to be a shadow here of this idea that there was also some unlawful scraping of the internet. That tends to be a new trend we're seeing is that copyright owners are not just arguing that they're downloading unlawfully, say from, you know, the BlitTorrent protocol through it, but rather they're saying, well, in addition to that, you're scraping public internet sources that we might have licensed, but you're using it in a manner that we didn't license you to use it for. So it's a bit more nuanced in that manner than it would have been previously. But certainly one might imagine that within that original settlement, some of this content might have been in there. It's kind of hard to tell, but so much of this is sort of shrouded in secrecy by Anthropic. And that last fact that he mentioned was actually referenced by the plaintiffs in the complaint. Quote, music publishers do not yet know the full extent of defendants unlawful torrenting of their works, nor all the ways in which defendants have exploited the works that they pirated in this manner because defendants continue to conceal the critical details of their infringement. Now, aside from just the IP theft issue, generative AI is also taking a growing market share when it comes to money that's generated through streaming services. Each time a song is streamed through apps like either Spotify or Pandora, the singers, the songwriters, and I guess the music creators, they earn royalty payments. And those royalties, they ultimately come from advertisements being served to the listeners. It's just the business model loop. AI generated music, however, has been flooding these apps while cashing in on the royalty payments. Deezer, for instance, reported that for the first time ever, AI generated tracks exceeded half of all the new uploads on its platform in the month of June 2026. AI generated tracks that cost essentially nothing to create can now compete for the same listener and the same royalty dollars as a song that might take a human artist months to write and record, an issue that Mr. Sigmund believes is the driving factor behind this lawsuit. One of the interesting dynamics to that, honestly, is that many folks are worried that eventually this material will get so good that it can really genuinely compete with music, movies, etc. We're already seeing that a little bit with the model Suno and a lot of just AI generated music popping on Spotify. So this is a very significant risk for these music companies. And this is probably one of the many ways in which we're trying to fight it back against what they perceive to be an existential threat against their entire industry, really. It's completely free to attend and it's going to feature some great guests like Bill O'Reilly, who I'm sure you know hosted the highest rated cable news show in history for over 20 years, as well as Emily Wilson from Emily Saves America. And in the webinar, they'll walk you through why the global financial system is under serious strain right now with record debt, a weakening dollar and inflation that's not going away. What you can actually do about it, they'll cover how gold and silver work as a hedge to diversify your portfolio with precious metals, as well as how you can move money into a gold IRA without paying a single dollar in taxes or penalties. And if you stay for the entire webinar, you're actually automatically going to be entered to win a free gold coin. Plus, qualified attendees can receive up to fifteen thousand dollars in free silver. And so check it out. You can sign up for the free webinar right now. It only takes like 30 seconds and you can find the link right there in the description box below. And one recent example shows, I guess you can say the relative ease with which these AI songs can be used to take money from legitimate music artists. In March, you had the DOJ announce that Mr. Michael Smith pled guilty to creating hundreds of thousands of AI generated songs and then using bots to stream them billions of times. Basically, in an extremely entrepreneurial manner, he cut the humans completely out of the equation. A robot, an AI robot, made the music and then a robot listened to the music. Also, that robot was served an ad before listening to the music, which was essentially defrauding the advertiser and therefore the platform as a whole. Prosecutors say that the scheme generated more than eight million dollars in fraudulent royalty payment. Here was our new director of national intelligence on this particular case. Quote, although the songs and listeners were fake, the millions of dollars Smith stole was real. Millions of dollars in royalties that Smith diverted from real deserving artists and right holders. Smith's brazen scheme is over and he stands convicted of a federal crime for his AI assisted fraud. Now, Mr. Smith has yet to be sentenced, but the wire fraud charge that he pled guilty to carries a maximum sentence of five years in prison. Mr. Smith also agreed to pay eight million dollars in forfeiture. Now, another issue facing musical artists is that generative AI can take recordings of an artist's voice and then produce brand new vocals that sound remarkably like that artist, almost eerily similar to them. Heart of My Sleeve, as an example, was a song written and produced by an anonymous TikTok user with the handle ghostwriter977 with AI vocals made to resemble Drake and The Weeknd. They were made using AI voice clones from those two artists. In 2023, that song blew up. It went viral and was briefly available on both Spotify as well as Apple Music before they ultimately got taken down due to concerns over copyright. However, in that brief window of time, it really did blow up. It was even considered for a Grammy for best rap song as well as for song of the year. Because the thinking went, it was written by a human who simply used AI to change the vocal. And the release of that single, it sparked really widespread and significant debate within the music industry regarding the ethical and legal implications of both AI generated music as well as vocal likeness rights. That debate then motivated some celebrities and some singers to try and trademark their own voices, which is exactly what Matthew McConaughey did when he trademarked his classic, all right, all right, all right, catchphrase, which I guess I said it, but don't sue me. I'm just a news reporter giving you the news. Taylor Swift also did something similar. In April of this year, her team filed trademark applications with the U.S. Patent and Trademark Office for two audio clips of hers as well as one image of herself. They're basically patenting Taylor Swift's voice and her image. And it was essentially a calculated legal maneuver, a preemptive legal maneuver to protect her voice and her likeness from imitations of AI deepfake generated videos and music tracks that might become available sometime in the future. And it's a novel legal strategy and we'll have to wait and see if it actually works. It's yet to be tested in court. Also, it's not even clear whether this can be done, given the fact that she filed for it in April and now five months later in September, Taylor Swift's trademark for her voice and her image have yet to be granted. So we'll see where that goes. Now we asked Mr. Sigmund on his thoughts regarding celebrities trademarking themselves, their images and their voices. And here's what for those folks, I mean, needless to say, their respective movie and music catalogs are a big issue as well, right? Obviously you don't want people training on Taylor Swift's music. And she's mentioned in here a few times, at least in the news coverage on this case. But in addition to that, there's this whole different layer of, even if it's not your copyrighted material, your, you know, what we call your NIL, your name, image, likeness, it's such a valuable thing. That's a whole different issue with these models because they're so good at creating what people would call deepfakes, realistic depictions of a voice, et cetera. So those species of cases, while slightly different, it's all the same concern of you've got a lot of data, you've got a lot of my stuff. And if I don't do something right now, I'm afraid that you were going to either, you know, a surreptitiously take away my ability to get licensing for other people. Like in the lyrics case, B you're going to learn how to compete with me, which is a concern or C you're going to use my good name to go market some random crap. And I don't like that either. Now at the federal level, no AI voice cloning law exists, but at the more local level, you have several States taking concrete actions to protect the intellectual property rights of the artists living in those particular States. For instance, in Tennessee, in 2024, they passed something known as the Elvis act, which gives people legal property rights to their voice, which protects artists in the state of Tennessee against unauthorized AI voice cloning. Now Elvis in Elvis law, it's an acronym standing for ensuring likeness voice and image security. Now, before the Elvis act, Tennessee's right of publicity law, it protected a person's name, their photograph, as well as their likeness. But the new law, it explicitly adds voice as a protected property right as well. However, this law, it only is geared towards someone trying to monetize someone else's voice through AI by claiming it's real, such as by, for instance, creating a song with the voice of a Tennessee based artist and making it look like it's really them. The law very specifically does not apply to using someone else's voice or likeness for social commentary or parody or comedy or something like that. And Tennessee it's worth mentioning is not alone. Eight other States follow their suit with their own laws akin to this Elvis act. Although Pennsylvania is a little bit different in that their law makes it a criminal offense to use an AI voice clone to defraud or injure someone. So Now in 2024, there was a study commissioned by SISAC and it's cited as the first ever global study of the economic impact of the use of AI in the music industry. Now, for your reference, SISAC is also an acronym and it is the international confederation of societies of authors and composers. And basically it's like a nonprofit NGO protecting the rights of global creators, which includes artists, musical artists. And for their study, after analyzing the current trajectory, SISAC wrote that AI generated music could represent approximately 20% of the revenue of traditional music streaming platforms within the next two years with an estimated market size of over $16 billion by 2028. While at the same time, potentially putting 24% of music creators revenue at risk, which would be over $10 billion. However, of course, if you are seeing the trend and you're good with computers and you have kind of a knack for selecting what works and what doesn't, hey, they're telling you that in the next two years, that is a growing sector of the market. However, if AI ends up dominating the music industry, if they're correct, and this trend continues, it'll have to be interesting to see what it ultimately does to things like live events, concerts, and music festivals enjoyed by millions. If you have such a huge sector or segment of the popular music being created by AI, purely using AI tools, what will people actually go to see live? I mean, could those even continue? I guess for a certain segment, you'll always have live artists, so those will continue. But what would people even be going to pay for? What would they be paying to go see with an AI artist? Conversely, though, you might have every concert just be like a guerrilla style situation where you have an animated avatar on the screen performing the music on stage. So you'll have to be interesting to see what will happen. Also, inevitably, AI-generated content will try to copyright-protect their own material, and that will be the ultimate ironic twist of fate, where you have the AI creators trying to copyright their own AI material so that other people aren't able to steal it from them. So a very fascinating turn of events. And basically, I would say that this is as much of an existential threat to the music industry as what happened in the 2000s with Morpheus and Kaza and Torrent websites beginning allowing people to, and Napster, allowing people to download music illegally. That was an existential moment for the music industry. Now, of course, they bounced back and you had the rise of Spotify and Pandora and Apple Music and things like this. But this is a new shift, right, of AI-generated music, and some of that stuff is crazy well done. In the last, I don't know, two months, I was listening to music on YouTube, and as you're listening to one song, it automatically feeds into another song. And I would say maybe seven, eight times I would be listening to some music, and I'll be like, man, this is so good. It kind of digs into your brain and it really hits the right feeling and the right vibe and the right notes. And then I try to find out who the artist is, and I can't. And I realize, oh, this is just AI-generated music. There's literally no artist behind it. There's just a creator, I guess, that's using the AI. But it's like the vocals, the synthesizing, the production, it's like all there. So I mean, I can only imagine by the year 2028 how good AI-generated music will be. So anyway, that's where we're currently at. Leave your thoughts in the comments on what you think regarding this development in the music industry. And then smash those like and subscribe buttons if you haven't already. And then until next time, I'm your host Roman from The Epoch Times. Stay informed, and most importantly, stay free.