RealmIQ: SESSIONS with Peter Csathy

FULL EPISODE: In the RealmIQ: Sessions podcast, hosted by Curt Doty, the guest Peter Csathy, a leading expert in media, entertainment, AI, and technology, discusses aspects of AI and its impact on the creative industry. Csathy shares insights on copyright infringement issues related to AI, the role of legal frameworks in the evolving AI landscape, and practical advice for artists and creators.

Big Topics

1. Introduction to Peter Csathy and his Background
2. AI and Copyright Infringement
3. Impact of AI on Media and Entertainment
4. Practical Use of AI for Creators
5. Case Studies and Examples
6. Future of AI in Creative Industries

Quotes from Peter Csathy

"The mere fact that you are scraping my copyrighted work without consent or compensation, that is infringement. That is an infringement that requires damages to be paid for that harm."

"If it's noncommercial, it's for education purposes, you know, things like that. I wouldn't worry too much about using generative AI."

"Like the last John Lennon song that McCartney used generative AI to bring to life. None of us would have heard that song unless generative AI was there to make it sonically possible."

"You see a number of companies now that are creating these ethically sourced LLMs or generative AI engines or platforms."

"I think that blockchain and smart contracts... it's a perfect match. That's a use case for web three that makes a lot of sense."

"There's going to be some kind of compensation system that operates for everybody else. And I'm working on one such system with some people right now."

Quotes from Curt Doty

"I think it was a big risk and a big leap for Toys R Us to say, 'We did this all with Sora. Aren't we cool?'"

"I think the magic bullet is to not say this whole thing was created with AI. We need to move towards the efficacy of the technology that it helps and empowers people who have these jobs, these creative jobs, to do their job better and more efficiently."

"I believe that creative people will be solving these problems. I don't think the lawyers are going to solve the problems. They'll execute the contracts."

"It would be interesting to have AI and blockchain get together and figure out some of these problems."

"I'm excited about those exciting events, those milestones, creative milestones that transform people's interpretation of what looks like controversy to actually, no, it's really cool."

Peter’s Podcasts
Major music focused podcast on the Consequence podcast network called “The Story Behind the Song” – where I interview the most iconic artists from the last several decades about their most iconic songs - https://consequence.net/category/consequence-podcast-network/the-story-behind-the-song/

“the brAIn” weekly podcast – my companion to my weekly newsletter of the same name - https://themediabra

Transcript

Hi, I'm Curt Doty with RealmIQ. This is our podcast, RealmIQ: Sessions, where we talk about everything AI with AI leaders from around the world. Please give us a follow or subscribe. Today's guest is Peter Csathy, who is a leading media entertainment, AI and tech expert. It was Worn about every hat in the industry in his 30 plus year career that spans traditional and tech driven media.

He's a frequent speaker and writes a weekly column about media, entertainment, AI, and tech on The Wrap. And he is a fellow podcaster. He is also a graduate of Harvard Law School and began his career as an entertainment intellectual property attorney. Welcome, Peter. You are the first lawyer that I've had on my podcast.

So very honored to have you. And we also share a past at Universal Studios. Yes. Thanks, Curt. Thanks for having me on. And just so everybody understands out there. Started off as an IP entertainment lawyer, and then I became general counsel of one of the operating divisions of Universal. But since that time, I've really been a serial entrepreneur, even though I still work on some legal matters for my clients, but nonetheless, primarily it's really on the business advisory side. 

That's cool. Wonderful perspective. And in terms of through the lens of A. I. Yeah, the lens of legal. I think we want to get into certainly your point of view on what's happening. What's happened with A. I. around copyright infringement. Yeah. Myself, I am an optimist. I know there were dirty deeds done that may be worked out in courts.

By the way, done dirt cheap. And we'll get into that. Yeah. Dirty deeds have been done dirty.  Yeah. Well, the internet's free, right? It's there. It exists. It's to see it's there that you put it up. We can take it. Right. That's the mentality free for all. Anyway, So let's get into that. So, for our listeners who maybe haven't delved as deep into copyright, yeah, have what is going on?

What's your stance and what are we facing in the, in the near term and the long term?  Very good. So having said the fact that I'm not primarily a lawyer now, what's important is that background, because I do follow it very closely and what I would say to everybody out there. Your eyes may glaze over when you hear the words law and legal, but it's truly like all these issues are existing.

It's all business. It's all business issues. It's all about where the money goes when value is transferred. So, no matter what role you play in the creative ecosystem or in the tech ecosystem, in AI it all matters. So, it matters to you as an artist, it matters to you as a producer, as a creator, as a lawyer, as a, you know, venture capitalist, everything, as a, as an entrepreneur on the tech side.

So that's why this is a really important discussion, Curt, and I think that's the most important thing, one of the most important things just to kick things off, that legal doesn't mean that we get into the weeds about it, but it's an important thing to track and understand. So that's, that's that. When we, when I talk to people about this. I certainly have strong perspectives that I'll get into but I always say that I come at this stoically,  passionate opinions. I come at it stoically, meaning that the world is as it is generative AI is here and moving fast. It will not go away. We keep for those of you who want to wish it away. It will not go away.

It will only speed up in terms of its progress. So, what's critical is you can only control what you can control. What you can control is how you act in that reality. And so, acting in that reality, I would say to everybody out there is to stay on top of these issues, but to not fear general generative AI, but somehow find a way to bring it into your life.

Learn about it, learn the jargon of it. That's really important. And then experiment with it, whether you utilize it or not. So that's, that's the kind of the high level to everybody as a long preface. Now, having said all that, these are the issues to answer your question directly.  So, I kind of started things off with, with, with generative AI, you have a. 

Massive amounts of value that have already been created in the marketplace. That's why NVIDIA has leapt into 3 billion or a trillion dollar plus. I don't know where it is today because the market's crazy, but it's somewhere hovering around that. And that's because of generative AI. So, value has already been created extensively.

And then you see Microsoft has really. Catapulted itself in the last couple of years because the generative AI to now be one of the sexiest companies in the world after being really kind of a sleeper for quite some time. And that's because, of course, Microsoft essentially owns an AI, which is seen as being the leader.

Microsoft is the, you know, leading investor and so, for all intents and purposes, many of us see open AI as being really controlled by Microsoft. Microsoft Mechanics So the copyright issues that come into play matter because it's about this transfer of value. How is the value created? So generative AI is a great, there's a powerful new technology, but it's a technology that has no power, no utility, unless it was able to train on data.

And that's where we get into copyright because what is that data that it trains upon? Well, as it relates to media and entertainment and everybody in the creative community, and then beyond that, that data includes. Millions and millions of the world of copyrighted work that find their way onto the internet.

So yes, on the internet, but they're copyrighted work. That doesn't mean that they can simply be used for whatever purposes. That somebody wants to do like if it's merely because a book is available online, I can't copy that book and put my name on it and republish it, you know, things like that. So, there's an example.

So in this particular case, what we have is we have a number of generative AI companies, open AI, Google, you know, et cetera, et cetera.  Anthropic and Suno and UDO will get into on the music side where they train on Different data sources, but they all say three words  Available data. Okay. Can you repeat that again?

You bleeped out there for a second? Okay Three key words, publicly available data.  Anytime you ask these companies what their training data set is based on, they will say publicly available data. And then when you ask them, well, what does that mean? Does that include copyrighted words? Then they'll all say, well, that's.

That's our special sauce. We can't tell you what's in there because that's what differentiates us. But it's been demonstrably proven, including just recently by Wired Magazine, that as an example, you have open AI has trained on at least hundreds of thousands of YouTube videos, as an example, and on and on. 

It's essentially, and with near certainty, all these generative AI companies. Have been using unlicensed copyrighted works to train their AI, which means to give them utility, which then created their trillions of dollars of value that have come on board. So those copyrighted works, of course, are from writers like you and me from studios out there from artists out there, musicians out there, whatever, and none of this content has been licensed.

So, the copyright owners see this reality and they say, Hey, hey, man. You can't, you can't just take my, my works, my copywriter works and put it into your sausage factory that will spit out some kind of generative response. You can't do that. You're, you're taking all of this data without paying me for it.

And you're clearly creating value for yourself because look at your market capitalization.  And then part of that, of course, is, well, you're taking this without paying me. And you're frequently trying, almost always trying to create a market substitute for the content I'm giving you. And so that's the poor argument from the copyright owners is that they, that they're, they contend.

That just taking, scraping all of this copyrighted content without, without consent is itself copyright infringement. Now on the other side, the big tech companies and the generative AI companies, they'll say, hold, hold on, man, you know, hold on.  That's fair use. We are only taking what's available, what's out there in the open internet, and we are, we are creating something entirely different in value.

So, it's, it's a fair use that we're creating, which is, you know, fair use is really a kind of a public good concept.  That fair use is created as a defense to copyright infringement. As a public policy thing that because it benefits overall, it benefits society. So that's kind of the core to it.  And they will say these generative AI companies will say it's fair use.

We're creating something newly powerful for all of society. And. We're not trying to create a market substitute for your content. We're just creating something that itself generates something entirely novel that doesn't compete with you. Derivative, derivative, transformative, or kind of words. And, and so, and it's not substantially similar to, you know, what your work was.

And if you don't mind me, bloviating a couple on about this, So that's the core thing, but one critical piece for everybody to understand there, after I've kind of set the stage, is that there are two pieces to the copyright infringement analysis when it comes to generative AI and copyrighted works. And that's frequently compressed into one.

At least that's what generative AI companies typically want to do. So, they want to have a confluence of the two rather than separate.  There's the input side of the equation. So, taking the copyrighted works, putting into the sausage factory so that the generative AI can train. So, on the training side, that's the input side.

And then the output side is what comes out. Once you make a prompt, what is the output that's created by the generative AI? And is that copyright infringement? And so, on the first piece, the infringement piece for training there, there's no requirement for substantial similarity. Okay. That under the law, it's just a novel application of existing precedents.

And we can talk about that, Curt, if you want to. So that's one issue, one side of the equation. And copyright owners, and I would be one of them, would say, the mere fact that you are scraping my copyrighted work without consent or compensation, that is infringement. That is an infringement that, that is pay, requires damages to be paid for that harm.

And then the second piece, the output side is a difference analysis, whether the output from a prompt is substantially similar to the input that was there. So, I'll just give you one, one specific example.  So, like the New York Times, and we'll get into that case more. One of the probably the most important. 

is the New York Times sued OpenAI and Microsoft for copyright infringement. And in their complaint, which is currently winding its way through the lower court, it's both the input side, it's saying that you scraped all our works without consent, and by the way, those works are behind a paywall. And so that's infringement.

But there's also the, they have proven by their own searching and their own use of open AI, that the output, the actual spitting out the generated response, sometimes even verbatim captures some of the articles that were scraped, but frequently does concise analysis of these things, including a feature stories that cost hundreds of thousands of dollars, if not millions, sometimes. 

So that's a, a lot there occurred, but yeah, that's, that's the stage. I appreciate that. And it is complex just speaking to the output side.  Yeah, because you know, I output imagery and I know I can never own the copyright of the imagery that comes out of Dolly or mid journey. So, so if nothing coming out on the output side is actually able to be copyrighted. I mean, there is a legal stance recognition that there's, there's not enough of that human origin to determine whether it's copyrightable. Is that correct?  Well, that that question about how much human interactivity is necessary for the output to be copyrightable is not black and white that hasn't like there is no distinct standard of other than the fact that the copyright office has said that there needs to be, I forgot the exact word, but some meaningful contribution from a human for something to be copyrightable.

Yeah. And so, it's like, this is again, just like so many things in the law. It's shades of gray rather than black and white, but certainly in the litigation right now, such as New York times, they are saying that it's an infringement because it's a market substitution that open AI is trying to create by scraping its works and then generating responses that are in the world of news and gathering, whatever that they've invested millions of on, and they're doing it. 

You know, open AI, Microsoft and the others are creating these. Full search capabilities to compete with and replace the New York Times is and so like taking all that value. So, it's infringement on the output side as well. That's what they're doing. Yeah, and obviously the newspapers are having a hard time anyway.

Anyway yeah, at the dot coms, you know, seem to be doing okay. Once they put up the paywall, right and help save the industry from selling newspapers, which. That's a dying industry, unfortunately.  So, so in terms of how an artist, a writer  who wants to use AI what are the, some of the cautionary tales that you can give them  with the context that you know you're also, you also use it, but you, you must be using it with caution because you know, the repercussions.

So, some of those guidelines.  Well, look, you know, I think that the, the law certainly finds between individuals who are using  potentially infringing services based on whether something's a personal use, whether something's an education use, a nonprofit use, or a commercial use. You know, there are different gradations there.

So, if something becomes more commercial, it becomes more troublesome for sure. But individuals are going to be in a different boat Then large-scale companies that are using it for more significant scaled purposes, and then there's another order of magnitude for companies like the platforms themselves, which are creating these enabling platforms for billions of us to be able to generate responses.

So, it's kind of a great gradation there. On the individual side, if it's noncommercial. Then, and I'm just giving you here practical thoughts for people out there. If it's noncommercial it's for education purposes, you know, things like that. I wouldn't worry too much about using. You know, generative AI, like chat, GPT or whatever, or mid journey or something to create an image or, you know, Google Gemini to create an image or something like that, I wouldn't worry to be like for a newsletter.

That's a non-charged newsletter, a free newsletter. I wouldn't worry too much, too much about it. Done a lot of that, but yeah, okay. Yeah, yeah, exactly. So, I think you should be fine. Like there's going to be much bigger fish to fry, but here's the thing.  And I wrote an article recently about this, like Toys R Us was just the first major brand to use OpenAI Sora to create a fully generative AI created commercial and, you know, obviously for commercial purposes and it made a big flash and that's one of the reasons why I did it. 

Po Ross was having a coming out party saying, Hey, we're still here. We're more vibrant than ever before. We're really tech savvy. Now. We're cool. And here, as an example, is this new, so produced commercial. And they, you know, you, it did kind of suck though. Yeah, I know. A lot of people thought that it sucked. 

It was really, it was first launching a can lion. And the reaction was mixed, but I would say more on the downside, but let's put that aside, but the fact that it was doing it to make a big splash, and everybody covered like everybody around the world covered. It was clearly for commercial purposes.  This, I would say, and I, my article said this, that I think that there is real risk for a Toys R Us.

Who is certainly savvy. They have lawyers and they're making such a big splash about it. They know the controversy that's going around, whether these platforms are copyright infringing or not, including open AI Sora. So I would say that there is still, there is a concept called secondary liability and vicarious liability where, or contributory liability that even though Poison Ross did not create the infringing platform, if it is.

Use it for commercial purposes at scale. I think an argument can be made that they also could be found liable for copyright infringement. Now, I'm not saying that's the way it will go, but there's risks there. And so, you asked a question, like, how should people think about it? The fact that there's risk in using these tools for commercial purposes at scale, that is holding back the industry, like the media and entertainment business, the studios. 

Madison Avenue brand. It's absolutely wind in their face, like resisting their use. Like fully embracing these tools because they're afraid of liability, potential risk of liability, and they're also afraid, understandably, of creative industry blowback for using this while all these issues are swirling in space. 

Well, What's interesting is there's an approach to say, this whole spot was done on Sora. Okay. The other tactic is going down a level of integration and, and You know, do a regular production, but then use AI and the tools that editors you usually use now Adobe Premiere has AI features. If you have to do some wire removal, rig removal, rotoscoping, special effects, lengthening clips, all these very utilitarian tools that editors have normally used, but now they're enhanced with AI.

Okay.  There's probably no copyright infringement because they're just,  you know, they're just using the information from that scene to like, well, I'm going to just make it longer and make it, you know, so there's just all these little baby steps that help make Commercials and or films more efficient and post production, you know, the phrase used to be, well, let's fix it in post, but now an editor can say, well, no, we're going to, we're going to fix it in AI,  but  they don't create the whole thing in AI, they're just using tools to help their normal job.

And I think it was a big risk and a big lead for Toys R Us to say. We did this all with Sora. Aren't we cool? And, you know, I, I think what needs to happen is that there needs to be this infusion of you know, building from the middle, you know, all the normal practitioners in that commercial process and venture incorporate the utility of AI that it offers them in their traditional uh,  Technologies, you know, whether it's Microsoft or editing or Google, you know, whatever tools they normally use there, those are now infused with AI.

And if they just do that and view AI into the process, but it's not the. The golden carrot say we produced this whole thing in ai Yeah, they mitigate their risk,  but they will have used it and those are steps forward in terms of adoption I think the toys r us thing was a debacle because  It was awful.

So, it's not going to fire anyone to say. Yeah, I just want to do those types of commercials It's like no one. Yeah, no creative agency wants to do that Creatives have their own creatives You know so it's you know, I, I think,  you know, the hype bubble around AI can actually complicate things more when it's used in the wrong way.

And I think the, the magic bullet is to not say this whole thing was created with ai. I, I think.  You know, we need to move towards the efficacy of the technology that it helps and empowers people who have these jobs, these creative jobs to do their job better and more efficiently. And they just happen to use it in their daily life as a commercial editor.

Let's say yeah, it's no big deal and there's no risk. And so, you know, that's a proper use. I think of a I that has no risk.  Yeah, I mean, look, you know, there's a number of layers there, so I want to be very clear  copyright and where I'm certainly sympathetic to the creator side when it comes to the question of where the money goes when value is created, that there needs to be a sharing of the money, that's, that's the central point that it can't be just enabling to big tech and generative AI to create all this value for themselves and not share with the creative community on which their IP trained or on their own.

Right. Technology train because it would be valueless without the content in the data, but to be very clear, I'm a huge believer in the upsides of generative AI, too. So, this compensation issue will be solved. And I have some ideas there, but it will absolutely be solved. Everybody wants to solve it.  Not every generative AI player, but most want to solve it.

And there's significant negotiations ongoing. Right now, with big tech and big media trying to figure these things out. Some have worked out as we know, and some haven't yet. I want to talk about one, a couple of that are really interesting, but then to your point about just tools and using them as different tools, like Adobe Firefly, which is the generative AI, which is built into the creative suite.

They say that. It's fully ethically trained. And so it was, you know Adobe trained it's AI only on licensed work that was already in its repository. So, when you're using that as a creative for marketing purposes or whatever, you should be entirely safe. I mean, there's some technical exceptions that they were busted on, but nonetheless, you should be entirely safe.

And so, what you see though, in the marketplace. If you have a number of companies now that are creating these ethically sourced LLMs or generative AI engines or platforms. And so, you, the user know that the only thing that's been trained on are licensed works. And so there is no downside risk. And then ultimately, of course, in whatever generative AI tool you use, there's going to be.

Most likely representations and warranties and indemnifications. So, like, if you're using generative AI in your big studio or whoever you, you are, then you have a contract with the, the big tech company that gives you, represents and warrants that there's no infringement, and if there's found to be infringing, then they will indemnify the studio for its use.

That doesn't mean the studio is not infringing itself. Contributorily, as I was saying, but at least you're indemnified for damages. So, there's that part of it too. I want  to give like another example that I think is really, a really cool thing because, you know, like you said, if you're an editor, if you're a creator and you, like you, You have certain nuggets of the film itself, but you just want to get more iterations of it faster so you can create like that's a really cool like previsualization or VFX or something.

Another use case that I think is really cool is that there’s all this content that's been created over decades, whether it's music, whether it's film, television, whatever made books, even what I don't know.  And it's sitting there collecting dust for from former lives. Now generative AI can be used to bring those to life, to reimagine them, to resurrect them.

So, taking black and white films and making them beautiful again, I mean, look at the example that like the last John Lennon song that McCartney used generative AI to bring to life. None of us would have heard that song unless generative AI was there to make it sonically possible. Like that's a really cool thing.

And I think that like, that's an example that really excites me.  Yeah. I mean, that's I call that innovation for good. But it's also creative concept, right? It's some, somebody creatively trying to figure out a creative solution to the age old question of, you know, what if John Lennon wrote a song that was never recorded, but we have is the lyrics.

We had the words; we had the poem. What would that sound like? I mean, it's an interesting premise, right? And that's somebody creative.  Saying that, and I think it's the creative community  that's going to harness these tools and take it  to a new platform, a new level that won't be done by, you know, who I call the hackers and just people who, you know, call themselves AI filmmakers when they've never done anything  a year ago that had anything related to do with film or didn't go to film school or never done anything like that.

So, you're right. I'm excited about those exciting events, those milestones, creative milestones that transform people's interpretation of what looks like controversy.  Actually, no, it's really cool. And guess what the estate of John Lennon approved it. So there’s no legal issues and it was cool and memorable and people know it.

And it's a great use of, you know, this technology.  Yeah, I mean, I get another example. Just recent example was there were several estates to your estate point like Judy Garland, where the estate gave permission for 11 labs, which is an audio generator to a voice generator so that now people could use Judy Garland voice.

It's been consented. It's been licensed by the estate. They can use it. And then the estate, of course, is compensated for that. So that kind of a use, like there's all kinds now, whether Judy Garland would love that or not, who knows, but nonetheless, it keeps like, here are ways to extend the legacy of people and bring them into new generations, which is cool.

I'll give you another example, if you don't mind, just because this is one that I also think is immensely powerful.  You look at artists.  Innovative musical artist, Grimes, great music, by the way. She and Musk were married at some point. They have kids together. In any event, Grimes is this great, innovative artist.

Soon after Chachi PT launched, she created a platform called Elf Tech, elf. tech. And what she did was that she gave music stems. So, like the little pieces of her music, she made them available as well as her voice. She made her voice available and stems of her music available to her fans and her fans could go to town on it.

She said, do whatever you want with it. Her only requirement is that if a song they, the fans create. Based on those pieces. If it is commercially successful, then she gets to share 50 percent of the revenues that are generated by the song. And I think that the whole idea of artists to be able to give their fans  Elements of their work, like almost like building blocks upon which the fans themselves could play with, build, whatever.

Yeah, that's really exciting. And then the artist, because the artists are giving the consent, they're setting the terms of what that looks like. They're in control. They're in control. That's which is, that's great for established artists and established franchises for those who are emerging and don't have big fan bases. 

That's not as easy of an opportunity. Yeah, but again Grimes a creative who came up with a concept using technology Empowering herself to monetize her content with her fan base. I mean these are these are And ways around these controversies and it's just creatively thinking out of the box,  you know, I believe, you know coming from creators creative people solving these problems.

I don't think the lawyers are going to solve the problems. They'll execute the contracts. Yeah, exactly. But it's going to be creative people saying, well, why don't we do this? You know, why can we do this? And you know, I, you know, so I think it's really important to the creative community who has a fear factor for, for, for some good reasons, but based on the very vocal minority of a few, they're, they're scared to get into the technology when there are a minority who are actively trying to solve these things, either through licensing or whatever.

And so, I think.  The more examples like you listed, the more those stories get out, that I think that it will hopefully ease the fear and help thaw what is really the ice age in Hollywood. Yeah.  Yeah. I mean, look, I’ll tell you something that I'm in the middle of. It's about solving that compensation problem.

So, this is the way I see it ultimately playing out. You have all these direct licenses that are being negotiated right now or trying to one of the most. One of the biggest ones so far has been News Corp license and entered into a license with open a I for five years. 250 million paid over five years.

That sounds like a huge number. My bet in five year’s time is that that will be seen as a deal that was very beneficial to open a I not so much to News Corp. That's my bet that it will be as great as it sounds today. be seen as a fraction of how great it was today, that's because of how fast things will grow. 

So, you have a bunch of companies that are entering into licenses, including the Atlantic. And there's some real good tidbits from the CEO who just recently said what the terms of that deal are with open AI. But then you have those like New York times who are, who said, no, what you're offering me isn't enough.

And they're taking this a long way. Now, how long they'll take it, who knows, but the more precedent that's set, the more deals that enter into play, that creates kind of like the baseline, the thing from the New York times, they're doing yeoman's work on behalf of the creative community to increase that number by putting pressure on.

So, like that gets, I would say, closer to a fair attribution between contributions on the, you know, on the.  IP side, creative work side and technology, but there can only be so many direct licenses because as big as we think that these generative AI companies are, they still have a limited number of internal business development people, lawyers, deal makers, et cetera.

So I think there's going to be some kind of compensation system that Operates for everybody else. And I think and I'm working on one such system with some people right now. That's You know, you know my perspective and I think that it's very fair to the, the, the, those who want to participate on the, on the media side, they will be setting the terms that will, they will still need to be agreed to, of course, by the generative AI licensees.

But nonetheless, I'm working on a mass solution in that regard with some very, very Well resourced people and then I think that they're ultimately also for the longer tail There's going to be some kind of like automatic opt in system You want your content to be scrapable and then be able to get paid some money from it You know kind of like a content id system for YouTube or a royalty system like in Spotify,  right?

You know, it sounds like what you're working on it sounds like Blockchain might be part of the solution smart contracts I don't know if you can say that or not, but you know, the altruism of what web three and blockchain was supposed to be about was supposed to be doing exactly that. How do you track one piece of content as it goes through a hundred iterations and still reward the original creator knowing that let's just say a TikTok video you could do duets with it the person who made the duet out of it could get compensated, but the original creator could get compensated You know that that compensation isn't happening right now, but there's these models of Not even using AI, but just content, you know, which has been a problem for a long time of creators getting their content ripped off and not getting monetized.

But you know, for two years I had explored that with another tech company that didn't quite get off the ground. It was, you know, trying to do social media done right.  But they ran out of money, but you know, sad story there. But the idea that blockchain and smart contracts that can track something as it flows through the internet seems like there's a solution there and would really be, it would be interesting to have AI and blockchain get together and figure out some of these problems. 

Yeah, look, I, I think that that what you just said is. One of the likely ways that this is going to be trackable and payable in that kind of sense. There was another, there was a company I was talking to, and I forgot which one now that was doing similar things, but it wasn't a blockchain like solution.

It was something else, and I forgot exactly how they did it, but from what I understand from the experts, you know, I'm not the, I'm not the engineer, but I'm pretty deep on blockchain web three stuff. Yeah. I think that that's a perfect match. That's a use case. For web three that makes a lot of sense. Yes, please.

Thank you. It wasn't nfts and it wasn't crypto, but it's something else. And its digital rights management, which has long been an issue, you know, certainly I was in the home entertainment business at universal.  How do you control your movies before they're released? They're already on a DVD in hong kong being sold on the street and You know You That's piracy.

So how do you combat that? Same with music. How do you combat that? So digital rights management has been an issue for a very long time. But now it's amplified with you know, all this new AI technology. Let's talk about music just a little because Suno and some of these other programs. Yeah.

Pretty cool. They're pretty cool. What are your thoughts about where that's going and how that has not necessarily been ethically sourced?  Well, so just in the last two weeks, the RIAA just filed major litigation against BOSUNO and UDO on behalf of the major labels for copyright infringement for the reasons that I indicated already.

And I'll give you an example. So, like there was an example on a recent podcast, I think it was Hard Fork, where they, they, One of those two services fit out an ABBA song, ABBA's like Dancing Queen, that sounded just like ABBA. It was slightly different melody, but it sounded just like ABBA. And so, you know, the question is, how were those, how was Sooner UDEO able to do that?

And so, it's pretty clear that Only because it was able to suck in and train on, you know, the ABBA songs that were out there and the artists, et cetera. So, first of all, with, on the music, just like the other cases I was talking about, both of those services and others like it say that they train on publicly available data.

And now, like, like I said, it's pretty demonstrably true that, and, you know, many have verified that they're training on songs and recorded music out there, like I said, and then there's the examples on the output side.  And, and in that case, so there's major lawsuits against both those companies. And in one of those two cases, one of the investors, this is not a good fact for them.

It's the quintessential smoking gun. One of the two investors was publicly on record, like in an interview several months ago, saying that, you know, if you probably wouldn't have invested in the company, if, if. That company was out there trying to do label negotiations or license the music because you know That would be too costly or whatever like he said that on the record So it's clear that what they were trying to do is just create, you know Like entrepreneurs do and especially in the tech world just move fast move fast and then ask for forgiveness later Forgiveness isn't working too well right now.

They're being sued So I think that those two services are in real trouble because there's a lot of power going after them And, and, and then, sorry, I, there's a dog barking in the background, but, but yeah, hold on.  But, but apart from that, there's another, there's another that I think is interesting to check out.

And it's called, um, it's called gem, I believe, and gem AI and it's ethically sourced. So, it solves itself. As being ethically sourced just like I was saying before so that's its selling point So it's a music generator that does the same things as suno and yudio, but it says it does it differently Yeah, and that's good.

I think that's going to be a new trend. Yeah, and I think That is alt tech working big tech Made all these original sins, committed all the original sins. I don't know how you get that genie back in the bottle, but I think all these all-tech companies are popping up and doing the right thing from all this learning, knowing.

That they don't want to get sued. So how do we work around that? What do we do? What do we have to, you know, what do we build our LLM off of and is it ethically sourced and then use that as positioning to help dispel the fear? I think that's all good. That will increase adoption and make, make it readily available to many people.

So, I I'm encouraged by that.  By the way on that, it's Jen, J E N, Jen music. So, Jen AI, and you'll be able to find it out there. And so, it's worth checking out. I hear good things about that company.  Awesome. We're getting towards the end of our podcast here, but I wanted to give you a chance to plug whatever you want.

I know that you have some Digital Hollywood sessions coming up. Yep. And so, I'm moderating two panels on Digital Hollywood. So, Monday night at 6 p. m. Pacific, it goes live so you can stream it then and it's free on Digital Hollywood. So, you can just register at the Digital Hollywood site. It will be on demand later after that, that video, but it's talking about the generative AI video timeline.

So, it's experts in the industry, a panel I created. Where they talk about, okay, what are the practical applications today in Hollywood is the video side for generative AI, where will it be a year from now, three years from five years from now, and whether there will ever be feature films, essentially that feature film quality that are generated, generated by AI. 

And when just a couple of the panelists to show you how cool it is and why it's worthwhile. So, Andy Beach, who's the CTO of Microsoft's M& E group, Immediate Entertainment Group, he's one of the guests and like that's a great voice. And then on Wednesday, I think it's a 3-p. m.  Pacific time. I have a legal sort of discussion about the status of all the major cases, what they are.

And then also like how a media company should decide to litigate or not litigate. What are the ways you think about it? Like, what are the steps you take? And I have one of the world's, well, one of the nations foremost. IP litigators. Who's an amazing, I've seen him in action. His name is Chad Hummel. And then Dr.

Moran McTeer, who's this spokesperson for the human artistry campaign. We'll talk. And that's the organization that represents over 200 media and entertainment organizations. It's a consortium. She's the spokesperson who goes to Capitol Hill to speak on their behalf, et cetera. So, she's going to talk about the status of regulation, et cetera.

So that's, thanks for plugging that. I think for everybody out there, it's worthwhile. And then. If you're interested in how generative AI impacts media and entertainment, I write a weekly, a weekly column in The Wrap, yes, but I also have a weekly newsletter called The Brain, and it's themediabrain. substack.

com, so it's on Substack, it's free, it's weekly. I give my analysis and I give the top 10 headlines from the past week. And I have a companion podcast called the brain that I started now too. So, for those of you who want to listen and walk your dog’s instead of reading, it's there for you. Awesome.  All right.

Well, listen thanks so much. And I want to have you back on again because there's so much to talk about still. But for those viewers and listeners, thanks for tuning in and catch more of our realm IQ sessions on your favorite podcast platforms, like Spotify and YouTube, and please follow and smash that subscribe button.

Thanks everybody. Thanks, Peter. Thanks, Curt. Really good to see you. Thanks for having. RealmIQ,  book your corporate AI workshop today.  CurtDoty. co branding, marketing, and product development.