Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_01It's important to keep creating human-driven content. I mean, if hypothetically we're in a world where all text is written by AI, humans would kind of lose control of our language. And we're not really sure that's a good idea. So I think it's really important to keep having humans writing things and creating things. On the consumer level, I think consumers really want to know whether they're consuming content created by human or by AI. People just really want to know that, right? If they're reading a book and they find out the book is written by AI, people are very uncomfortable with that.
SPEAKER_00Hello, and welcome to The Media Co-Pilot. It's a podcast that explores how AI is changing media, journalism, and content creation. I'm Pete Paschel, longtime tech journalist and founder of The Media Co-Pilot, and I'm excited to bring you fascinating conversations with fellow journalists, media executives, and the people building how the business of news and content will work in the future. My guest today is Trip Adler. If that name rings a bell, it's because he co-founded Scribd, one of the first major digital publishing platforms that helped
writers and readers
SPEAKER_00connect long before Kindle Unlimited or Substack came along. Now Tripp is turning his attention to one of the biggest unresolved issues in the AI era, how creators get paid. Not to mention maintain control when their work is used to train large language models. Tripp's new company, Created by Humans, is building a marketplace where authors can license their work directly to AI companies. It's a novel approach to attention that you could say is definitely growing, and it's one that sits right between copyright law,
SPEAKER_00technology, and the economics of creativity. If it's successful, Creating by Humans could become a cornerstone of what some are calling the new creator-rights stack, the emerging infrastructure that will define how human content flows into AI systems. We talk about his vision for a fairer AI ecosystem, what he calls the fourth law of robotics, and how he interprets the recent anthropic settlement and the ongoing lawsuits between creators and AI companies. Most importantly, we explore whether there's a sustainable middle ground between
protecting creators
SPEAKER_00and fueling innovation. I hope you enjoyed the conversation. And if you do, I'd be grateful if you'd leave a review for this podcast. Reviews, especially in Apple and Spotify, really do help the show. Please leave one if you can. Please and thank you. Two more quick notes. One, I'm offering a monthly class on AI for beginners. It's only 60 minutes, and like everything I do, it's focused on practical takeaways for media professionals, people in journalism, public relations, or communications. And it's only $49. Each one's a live class,
SPEAKER_00so it's always up to date, and I guarantee you'll walk away with something you can apply to your day-to-day. You can sign up at mediacopilot.ai. Also for deeper training on AI, I offer two different six-week courses that are starting in November. One's for journalists, the other's for PR and comms. They're loaded with lessons on things like deep research, how to automate workflows, and building your own AI agent. You can sign up for these, where else? MediaCopilot.ai. Okay, now it's time for my conversation with Tripp Adler,
SPEAKER_00co-founder and CEO of Created by Humans. Tripp, welcome to MediaCopilot. Thank you. It's great to be here. Awesome. It's great to have you here. I love that you're here and I love your background because you are obviously one of the co-founders of Scribd, which to me is home of the original content ingested on someone else's platform idea now. You're focused on that in sort of a completely different arena.
SPEAKER_00That's probably a little more complicated. So I can't really think of anyone better who would understand all the dimensions of what's happening with AI and content, and especially on the ground level. So yeah, tell us a little bit about that background, since listeners might not be super familiar with what you did with Scribd and you know, how that path led you to Created by Humans.
SPEAKER_01Yeah, sure. Yeah. So, yeah, I started Scribd when I was just out of college in 2007 and started as a way to publish and share documents on the web. Of course, there was all sorts of, you know, corporate issues that we faced there. So, quickly got a pretty good crash course in how copyright works and how the DMCA works. Through that, I got to know the book publishing community really well, started to really understand what that whole industry was about,
SPEAKER_01and then expanded Scribd to pioneer the subscription model for books. So we worked with publishers of all kinds to put that model in place and grew that up into a big business with It's a big scale business now with over a billion users, over 200 million revenue. And yeah, in the process, I got to know the book industry really well. I got to know the tech industry really well. And then when this AI situation came along,
SPEAKER_01it was kind of a new collision of the book industry, the tech industry, but in a new way. And all these books were trained on without... first checking in with the authors and the publishers, and it resulted in a bunch of lawsuits. And I actually had a lot of people in the book industry reaching out to me, asking me for my advice on what to do about this situation. I had people in Silicon Valley asking me questions about the book industry, and I just figured that this just seemed like the right thing the right problem to solve is to try to bring these two
worlds together and get a get a model that connects the book industry and the content industry more broadly in the industry and started created by humans try to solve this problem and
SPEAKER_01we're now we're making some good progress in solving it nice so tell me the central thrust of created by humans but also beyond just the sort of mission and why it even exists um where it kind of fits into this whole
SPEAKER_00sort of myriad web of um you know content and ai um there's a there's a lot of stuff that you might even say there's like a legal stack as an addition to sort of a software stack on this, you know, and it sort of depends on where, you know, which party you are, what the sort of precedent is. I don't know, there's a million sort of dimensions to look at it. We'll unpack as many of them as we can in the time we have together. But let's just start with the fundamentals.
SPEAKER_00What's Scribd doing? And then again, where does it fit?
SPEAKER_01You mean what Creative Behavons is doing?
SPEAKER_00Oh, sorry. Created by humans, not script. Yeah, confusing. Multiple companies here. No, that's me. That's just my age. I remember script. It doesn't matter. Go.
SPEAKER_01Yeah, so we started Created by Humans to try to solve this problem. When we started the company, we really did – we had no idea how to solve it. We just looked at it and we're like, well, this seems like a really tricky situation. And we're like, well, someone's got to sort this out because either copyright law is going to just disappear or – or copyright is going to hinder AI innovation. And we thought neither of those were kind of good solutions. So we started to just try to solve the problem and create a world where, you know, copyright is respected, you know, humans keep creating great things.
And then meanwhile, there's a business model for how that interacts with the world of AI.
SPEAKER_01And, you know, AI products can be built with copyrighted content and AI can thrive and, you know, and human generation can thrive and these two can kind of work together. So we started just trying to solve that problem. And we've been building a lot to solve this. It's just not one thing. But what we started with was kind of like the legal piece, which is building an inbound-outbound licensing model. Because when we first heard that the reason why these AI companies... didn't want to, um, or, you know, didn't just, just wanted to train everything without licensing. It's just, it was too complicated to license.
SPEAKER_01Cause there's, you know, there's millions of books out there. Um, in order to clear the AI rights, you have to, you have to get the publisher on board. If you get the authors and then there's multiple authors, there's illustrators, it gets very complicated. Um, So the first kind of approach to solving this is basically clearing these rights for these books. We had to define the AI rights and then start to clear them for books and then make that into a super easy package that AI companies can then license. And this way they can get the data they need and just make it
really easy for AI
SPEAKER_01companies, but also make it easy for creators and just make it easy for everyone overall.
SPEAKER_00So, okay, so the whole idea is that you're going to the publishers and then through them, advising them, getting them to, I suppose, adjust their contracts with the writers and getting their writers on board for these rights for AI to do stuff with their works, right? Yeah. and um that sounds like a big process in and of itself uh and but the the dimensions of like what ai can do with it i'm actually really
interested in because it's like you that uh can you sort of summarize how that breaks down because there's like training there's rag there's agents i don't know if this covered your
SPEAKER_00your contracts cover all of that but what What is like a that, I guess, section of these contracts? How is it sort of evolved? Like what is what can an AI do with one of these if it is cleared?
SPEAKER_01Yeah, so the first thing we did is we defined this concept of AI rights. When AI came along and it collided with the content world, there was no structure for how AI rights should work. So we defined the concept of AI rights, and we realized there are actually three types of AI rights. I mean, there might be more over time, but we've started by defining three that we think largely capture what needs to be captured. So pretty similar to what you just mentioned. So the first one is training rights. This is the right for a piece of content to be used by AI models for training.
The second is what we call reference rights, which is the ability to use content for RAG. We call it reference rights because it's kind of like more of a bookie word.
SPEAKER_01It's like the idea that these AI models can reference the book using RAG. And then the third category of AI rights, we call them transformative rights. And that's the ability to take a piece of content and transform it into something new using AI. Because in the future, you could take a book and you could personalize it with AI. You can adjust the length with AI. You can turn it into a movie with AI. I mean, you can do all sorts of things. So we're building what we call the transformative rights framework for that. So that's kind of how we've defined it so far.
SPEAKER_00Wow, that's interesting. I really want to sort of dive into that a bit. So as I understand, you've also kind of got this, I don't know how literal this is, but I was reading about you've defined sort of a fourth law of robotics or something sort of off of the Asimov laws. Is that... at a higher level than this sort of specific stuff we've just been talking about in terms of like what's guiding the principles here. Can I explain what that is? Like for those that don't know, the three laws of robotics are like the do no harm or don't harm humans.
Maybe you could explain like whether those apply and what this fourth law is.
SPEAKER_01Yeah, so we created this legal framework for AI rights, but we realized that to really solve this, we had to kind of get a little philosophical here. Because, you know, when we first started on this path, this argument is being made, and this argument is still being made, that since a human has the ability to read all the books they want and then communicate what they learn from those books, therefore, AI models should also be able to do the same thing. And, you know, this didn't really feel right to us. You know, we thought that, you know, I mean, human can only read and communicate so many books in their life.
SPEAKER_01An AI model can, in a matter of moments, read millions of books and then distribute that to the whole world. So we thought that we needed a new way of thinking about this. And yeah, we were inspired by Isaac Asimov. He created the Three Laws of Robotics. And he basically said, like, in the future, you're going to need certain rules to regulate the relationship between humans and robots. And what we realized that we need a fourth law here, which is a law that regulates how robots can learn from human behavior and human
SPEAKER_01creative output. And basically what the fourth law says is that if a robot wants to, or an AI model wants to learn from a human, such as like trading on their book, this should happen in a way that benefits the human and the robot. And there's five principles underpinning the fourth law. There's, you know, humans have the right to consent to their data being used. They have the right to control how data is used. They have the right to get compensated for it or negotiate compensation for it. They should get credit for it.
SPEAKER_01And then there should be continuous learning of how these, there should be like a pipeline where these robots can always be learning from the latest and greatest creations of humans. So that's what the fourth law is. We created this as kind of like a philosophical framework, and then we built the legal model of AI rights kind of around that framework.
SPEAKER_00Nice. Okay. And so that's all great on a high level. Then we've just sort of talked about a little bit and how it works on the sort of practical level of what are the ways AI can use the content. And I imagine you're starting with books, but you're sort of coming at this from a little bit of like what content agnostic or format agnostic, I would imagine.
SPEAKER_01Yeah, I mean, our vision is to do this for all content, all human data. We've already started dabbling with some other content types like music and medical papers and things like that. But we realized that that's a huge problem to solve, and we wanted to pick one type of content and really solve it for that first before we expand. So we've basically gotten laser focused on books, and we're trying to just solve this once and for all for books. And then once we do that, we can expand that to all kinds of content.
SPEAKER_00Yeah, and the framework's straightforward, and I think a lot of these sort of proposals around this tend to be like, oh, I get it. You can ingest it. It pays a certain amount to do that. And then if it's doing RAG, it's doing – there's something – there's fees associated with all this, and it's all kind of done automatically in the background once all those have been agreed to. What I wonder about is, like, how much, right? Like, what is a fair rate for – each one of these things, depending on what the book is and how much it's being,
SPEAKER_00um, how much is being tapped for any one of these things. I imagine training looks a lot different than RAG depending on, you know, the popularity of the book for number one or the information it has. So, um, like how, how is that, um, determined and, and what sort of counts as a fair rate these days, like particularly for training, does that even, um, Is that even meaningful anymore? I feel like the training argument has sort of evolved a lot. It used to be all about training. Now we all realize, oh, inference is actually probably going to be the thing going
forward.
SPEAKER_00How do you sort of apportion that? How do you figure out what is a fair rate for any of these particular actions that the AI is doing on the information? Yeah, I mean, that's the big question, and I don't think there's a simple answer to it. Well, I guess the simple answer is that we're going to let the market decide. By creating a framework where rights holders have the ability to decide what deals
SPEAKER_01they do, and AI companies can negotiate the deals they want to do, we think with each deal that gets done, you kind of gradually move toward market rates. And as over time there's more and more deals done, there'll just start to be kind of standardized rates around this. The numbers currently are all over the place. I mean, of course, some companies argue trading is fair use and they should trade nothing. We're doing one deal right now where the company is paying a very large amount per book. So it really just depends. There's kind of like a really wide range of thoughts on pricing.
SPEAKER_01I think the market will just determine it over time. The other thing is like, it also depends on the right. I mean, I think training versus, um, reference rag versus transformative, which can be very different models. Um, you know, typically training the, the way I companies like do is like, just, you know, pay once and have, you know, perpetual use of the content, something like, um, rag or transformative, there's more of like a variable component, um, going forward. With Transformator, for example, you're transforming a book into something new, which is kind of like another version of a book sale. So there's just very different models for each of these rights,
SPEAKER_01and we think we'll just let the market kind of determine over time.
SPEAKER_00Yeah, I remember – I forget the precise model, but I was talking with another CEO about – it's kind of similar. It was about music. And it was about if you wanted to sample someone's voice that they had put on this platform that you could sample. So say you're, I don't know, Drake, just because he's been already through this. And then you licensed the voice and then you sold AI music based around that voice.
There'd be obviously for someone like Drake, the commission would be massive.
SPEAKER_00It would be whatever, maybe even more than 50%. And that would be kind of a similar model here, I would imagine, for the transformative stuff. So someone takes your book and transforms it, and then obviously the rights holder is going to enjoy a large amount. I imagine they'd set it themselves on what that portion would be.
SPEAKER_01Yeah, exactly. I think that there's going to be different rates for different people. I mean, a best-selling book would be different from a book that's not as big of a seller. And it depends also on whether the book is being transformed and the entire book is being reproduced, or if it's a rag use case, an AI product might just be getting one sentence out of a book, in which case that's a totally different model. Um, so, you know, I guess we, we believe there's just, there's, there's so much value in books and, you know, books have always been used in this kind of linear way where
they're sort of like
SPEAKER_01read from beginning to end. I mean, it started to change a little bit with things like searching within books, but AI just creates like, you know, infinite possibilities for what can be done with books. And we're just trying to unlock that through this, this new rights framework, but do it in a way that's in accordance with the fourth law where, you know, the author is, is, you know, maintaining control and getting compensated and all of that.
SPEAKER_00So if a writer wants to join Created by Humans or wants to get involved, I mean, is it through the publisher? Can they do a self-serve thing? How does that work?
SPEAKER_01Yeah, so we work with authors and with publishers. So when we started on this path, we started talking to publishers. And what came up over and over is that these AI rights have never been defined. And if they're not defined, that means that they're owned by the author. or at least it's a shared right between the author and the publisher. So in order to unlock this opportunity, we do need to work directly with authors. So we have built a platform where authors can sign up directly. So it's, you know, it's self-serve, it's very easy to use. You know,
SPEAKER_01author can sign up, they verify their identity, they select the books they want to license, they select the AI rights they want to license, and then they sign an agreement which license those AI rights to created by humans. It's very self-serve and, you know, very easy to use. So if you're listening to this and you're an author, we would love for you to sign up and give it a try. And, you know, it's a way that you can start to take control of your AI rights rather than just kind of, you know, AI companies just sort of using your books in a way you know what's going on. This is actually where you can take control. You can get compensated and you can have some visibility in
what's going on.
SPEAKER_00So let's talk about the sort of compliance enforcement side of this, which is always kind of a bit, I don't know, elusive maybe. So you get probably less worried about like the back of the people who have AI companies that are Well, actually, let's talk about that. Let's talk about that first before we get to like the rogue operators. So how does this work on that side of things? And what interest have you seen from AI services and maybe even
frontier services
SPEAKER_00on like what interest they have in working with sort of a broker, I guess, a rights broker like created by humans?
SPEAKER_01Yeah, well, it really depends. I mean, there's certainly some AI companies that are just, you know, have the opinion they're just going to kind of train on whatever data they can access and it's fair use and they're not working very directly with creators. And there are some that really want to do this right and they want to make sure everything is, you know, has the full approval of creators. So it's really been a mix. And yeah, we're not just doing the licensing. We also are building... you know, the infrastructure here around monitoring, enforcement, analytics, all of that. So,
SPEAKER_01because, you know, a lot of content owners, they, you know, they are sort of distrustful of what can be pinched on with their content. And by kind of building like a platform that sits in between the rights holder and the AI company, we can kind of like develop more trust with rights holders and make them, you know, more willing to license and more willing to be more experimental with what is done with their content. So that's kind of a whole other piece of what we're building out is that infrastructure layer.
SPEAKER_00Nice. So maybe then just keeping with the sort of reality of AI companies and what they think and what we've seen in the last few months, I think everyone's kept up with some of the bigger lawsuits. And one of them was the Anthropic settlement that happened. I forget what that was, probably a couple of months ago now. And I'd like to understand how you view that or how it's sort of being also interpreted in not just the author community, but the publishing community.
So the ruling, as I understood it, there was kind of, it was kind of split.
SPEAKER_00It was kind of saying that like, I don't know if it was definitive, but it sort of implied that Anthropic had some ability to train on these books as long as they paid for them. And then there was a little big thing about the pirated books that was not okay, so it ended up being settled. That's very broad strokes, I know, but maybe you could sort of break down what is cause for maybe concern cause for optimism, um, in that ruling. Cause I think there's a bit of a, maybe misinterpretation.
SPEAKER_00You probably called him, certainly probably not thinking about it this way, but the, um, the idea that they could just buy a book, um, legitimately ingest it and then they could do whatever they want with it i don't think that's necessarily what the ruling said but i feel like there's a sort of a conventional wisdom that like is sort of emerging post that i know it's not the final word but uh tell first tell me what you thought about that lawsuit how it's affected what you've been able to do and then where that sort of legal fight need is going from here do you uh just on the sort of broad level
sorry that was a long
SPEAKER_00question
SPEAKER_01Oh, no, it makes sense. Yeah. So I think that the lawsuit has and the settlement has mostly just been about a bit about piracy from what I could tell. I think that the main the main point of that settlement was that the books were pirated and that the, you know, Anthropic use them without buying them. And that's copyright infringement. And there's a payment of $3,000 a book for the piracy of the books. So I think that makes sense. I think that's a good thing that it's being established that an
AI company,
SPEAKER_01just the way an individual consumer can't pirate a book, An AI company also can't pirate a book, right? If they want to work with a book, they at least need to pay for it. So I think that was kind of the resolution there. And I think it seems like a lot of AI companies have pirated books, from what I can tell. So it seems like there's probably gonna be a lot more of these piracy settlements going forward. So I think just that alone, there's probably gonna be a lot of money going from AI companies to authors and publishers.
SPEAKER_00The second question... Oh, sorry, go ahead. Yeah, well, the second question is, so let's say the AKB does buy the book. Can they train that for free, or is there an additional payment to train on the book?
SPEAKER_01And I think that's still very much unresolved. I think the judge in the case made public comments that that training is fair use. But I think that it's still very much unresolved. You know, a lot of people like to talk about this about whether training is fair use, like a binary question. I don't really think it's binary. I think there's a lot of gray area, and it really – It really just depends on when the training happens, what's done with that work. And I think there's going to be a lot of different legal outcomes because there's so many more lawsuits.
SPEAKER_01I mean, it's just the first of many to come. So I really think it's still early days in terms of this question of whether training is fair use. And I think the ultimate answer is it's not going to be black and white. There's going to be kind of a middle ground solution there.
SPEAKER_00Yeah, it's also like why I think the model of sort of separating it the way you have is important because information and training it can't like it informs the model broadly and then that information tends to come up whenever it's it's kind of needed and again i know i'm generalizing here but it might not even, in other words, the query might not even be about the book. You know what I mean? It may just come up. And so in a query where it's directly about the book, and again,
SPEAKER_00use whatever term you want, if it's another piece of work other than book, but the idea that like, oh, well now I know you want something very specific. The intent here is to get material and that maybe it's a much clearer line of why there should be compensation. And likewise transformative because it goes even further. So that just kind of seems like, okay, it's okay to parse it that way. And to your point, it's not monolithic. It's not like one and done. These are different acts, different intents.
SPEAKER_01Exactly. Yeah. So, so there's, it's one thing to like train a book and maybe, you know, use that book to learn the English language better. It's another, it's another thing to take that book and then, you know, reproduce, um, a pretty good summary of that book that's made available to consumers. Right. Um, we would argue that's more of a transformative use. So, um, so that's, so one of the things we're doing is, is kind of drawing those lines. We're defining what is a training use, what is a transformative use and, you know, creating like real licensing model around these different concepts and drawing clear lines so that we could start to,
SPEAKER_01to, to be more clear about this.
SPEAKER_00Gotcha. So, um, So how do you design a licensing model that isn't too granular in that sense to scale or too broad in protecting rights? How do you think about that in terms of like, where's the right balance? Yeah, I think that's really tricky, and that's what we're working on figuring out. I mean, to date, we've had these three buckets, training, reference, transformative.
SPEAKER_01I mean, I think ultimately transformative is going to break into more sub-rights, and we're starting to define that because there's a transformative summary right, there's a transformative chatbot right, there's a... transformative, you know, book to video, right? I mean, they're all going to be separate sub rights. And we're trying to find all this. So I mean, it's an ongoing thing. And I think each deal starts to establish a little more precedent. And over time, we'll start to have like a, you know, a broader picture or broader answer to it. But I think it's like that's, that's the work we're doing is we're,
SPEAKER_01you know, we're working on solving right now. And how do you deal with what I imagine you're encountering a lot of default
SPEAKER_00skepticism and, you know, frankly, probably antagonism from the creator side, a large portion of which feel like this AI is taking over, it's ingested things. They just want to lock everything down. I imagine there's a bit of education that you have to sort of do when you walk into one of these publisher meetings or even talking to authors broadly.
SPEAKER_01Yeah, I mean, certainly there's a lot of education and there is a lot of just kind of, you know, general anti-AI sentiment where people just like don't, their books have anything to do with AI. But the good news is there is like, you know, a number of folks out there who are really enthusiastic about AI. And we do have a lot of authors and publishers that we're working with who are really enthusiastic about this. And I think we have enough support from AI. enough authors and publishers that we can kind of get this model going. And I think once we prove it out, we can then, we'll get more folks on board.
SPEAKER_01So it just, you know, it just takes a matter of time. It's like with Scribd, you know, we started the description model. Everyone thought that was a crazy idea to subscribe for books, but we had, we managed to get, you know, a few publishers on board and then that, that led to more going on board. And then eventually kind of, it just became sort of like understandable people would subscribe for books. And I think we'll see it play out similarly here, where once we kind of prove it out with the first early adopters and then we can expand to, you know, more broadly from there.
SPEAKER_00And on the other side of things with the AI companies and, you know, how are incentives there changing with these court rulings that we were just talking about, but also with the sentiment you're talking about bubbling up, you know, surely like, you know, as these things stack up, maybe they're more incentivized to start making deals or is anything shifting their behavior? Like what is your sense on that area directionally?
SPEAKER_01Well, I think for a certain type of rights, things like reference rights, those deals are happening quite a bit. I mean, that's already been happening for a couple years, and there's more and more of these rag deals need to be done. Transformative deals, there's, I think, less just because it hasn't really been defined yet. But as we're defining it, I think we're seeing a lot of interest on that. I mean, we're probably seeing the most interest in transformative rights of all the different rights we have. So training is the one that's the big question. I mean, that's the most legally complex one. Unfortunately, no one even really wants to talk about it because it's so legally complex and
SPEAKER_01there's so many lawyers involved. So I think it's really unclear where it goes. I think this Anthropic case is a big step, but I think there's a lot more to play out here. Um, so I think, I think training is just, is the most up in the air. Um, but I think over time, just, you know, as there's more of these, um, more of these lawsuits get resolved as there's more and more deals done as, you know, as we are doing, it'll just over time start to work itself out.
SPEAKER_00And I'm just a bit of a tangent here, maybe, but I'm curious about AI authorship, because I know the Copyright Office has been a little bit, I don't know, widening, I guess, on the stance on this, on if something is not necessarily wholly written by AI, but at least partially AI. is that that that can be um that can be copyrighted depends on the case i don't know if you've kept up with that or seen any sort of test cases
with it but um how does that kind of a weird wild card factor in all this or is it just kind of a
SPEAKER_00footnote for right now um
SPEAKER_01I'm not sure. I would say I don't really have a good, you know, anything good to say about that or anything smart to say. It's a really complicated issue. Yeah, I'd be like, can you copyright something created by AI? I mean, that's a really, really hard question to answer. What I would say, though, is, you know, we are... We want to try to draw lines between human-driven content and AI content. I mean, we think that it's important to keep creating human-driven content. I mean, if hypothetically we're in a world where everything, all text is written by AI,
SPEAKER_01humans would kind of lose control of our language, and we're not really sure that's... I'm really sure that's a good idea. Um, so I think it's really important to keep having humans, um, you know, um, um, writing things and creating things. Um, and then also I think, um, on the consumer level, I think consumers really want to know, um, whether they're consuming content created by human or by human or by AI. Um, you know, if, if, um, um, yeah, people just really want to know that, right. If they're reading a book and they find out the book is written by AI, people are very uncomfortable with that. So, so we want to work to kind of draw these lines between
SPEAKER_01human-generated content and AI-generated content. There's also a lot of middle ground because I think in the future, right, you might read a book and like some of it's written by AI, some of it's written by humans. And we want to work to sort of like establish kind of standards around that so that there's just kind of clarity in terms of what you're reading or what you're consuming.
SPEAKER_00Yeah, it's actually, I'm curious, like, I know when I talk about AI written stuff, it sounds like, well, like, especially when it's books, right? Because like, we all like sitting down with a good book and something human written. That said, there's a lot of different kinds of books. You know, there's like a lot of nonfiction books. And a lot of how-to books and cooking books or whatever. And so you can see AI having more of a role in that sort of thing. You could also see the sort of rights that we were talking about,
SPEAKER_00the three buckets, being very different for a book like that versus a fiction book. Yeah. Talk about that a bit. How does this change between the different types of content that are putting out there? Are there different trends? Are there different things to think about? Different rates on the licensing? How does it look so far? Is it more just getting started to make sure that this framework is in place?
SPEAKER_01Yeah, I think it's just so early, we're still sorting it out. I mean, the interest in AI companies from different kinds of books is, it's really all over the place. I mean, there's interest in, you know, there's nonfiction books, there's interest in textbooks, there's interest in fiction, there's interest in cookbooks. I mean, there's just, it's all over the place. And yeah, I think there's just different nuances to all these different kinds of content. And I think, yeah, I think these are the kinds of things we're sorting out, right?
SPEAKER_01Um, cause like, you, you know, you, you, you talk about like a nonfiction book versus a fiction book. Um, and you know, you want to read something written by a human, but in some cases, you know, you might be okay reading something that's written by AI or partially written by AI. So as long as we're starting to like, you know, um, you know, kind of create clear standards around this and creating clarity with everyone involved, we think that's kind of how it starts to work itself out.
SPEAKER_00Yeah, I guess how I would sort of think about it is like, what's more vulnerable to AI in a sense, you know what I mean? Like, and maybe it's just different nuances to it. Because like with something like a nonfiction book, I can see it, the AI summarization actually being sort of a beneficial thing to people like, oh, I don't want to read the whole thing, but give me the 10 page version of this book or the two page version of this book, or even just some key guides that it has. Yeah. And having AI do that is valuable. And obviously there would be, again,
I think a system like yours that would sort of manage that.
SPEAKER_00Fiction book, I could still see some things like, you know, say that it's someone like the vast majority of authors who don't have movie deals or TV deals and are like, hey, I would love to see this scene rendered by something like a vo3 or a sora you know like but use the carry like show me this kid this scene is created by ai how does that you know like does your model account for all that like and and are you actively involved in talks like
these like the use cases i guess i mean absolutely yeah we're talking to companies
SPEAKER_01that want to you know turn books into choose-your-own-adventure books, or want to turn books into virtual worlds. I mean, there's just infinite things that can be done. And we're trying to build out the frameworks for this, right? Without these frameworks in place, it's too easy for these AI companies to just take the books and just kind of do whatever with them, like what's happening with Sora and these sort of things. But if we can build the framework and make it easy for these AI companies to use our framework and easy for creators to maintain control and
have that biz model in
SPEAKER_01place, that's kind of how we solve this across the board.
SPEAKER_00Nice. I imagine there's a lot of... startups that are in the stack that might want to sort of be also in there, not just like the big AI companies that want to do things that are very specific with regard to licensed content.
SPEAKER_01Yeah, I mean, we have a lot of interest in startups, and there's so many interesting things going on. And most startups, they want to license content in this the right way. It's just it's so hard to navigate the world of rights. It's hard to do as a big company. It's hard to do as a small company. And we're trying to make that really easy for everyone. And I mean, I'm hoping we can drive a lot of startup activity around books. I think that if we can make it really easy for startups to build products on top of books, there could be like kind of a book industry startup boom. And that's something that we're trying to get going here.
SPEAKER_00Yeah, I remember one of the first AI experiences that I saw sort of demonstrated was someone uploaded a book into an AI. It was like, now you can talk to the book as a sort of a corpus of content, just sort of demonstrating that ability. And that's just the most basic 101 level of what you can do. Exactly, yeah. And if we can find a way to unlock the value of the IP and protect the IP,
SPEAKER_01but then also make it something even greater with AI, I mean, it's just a huge opportunity. There's just so much to be done.
SPEAKER_00Yeah. So that said, let's project out that opportunity a bit. Let's look five years ahead. What does success look like if created by humans catches on in all the ways that you would like it to do? What does it look like for creators? What does it look like for publishers? What does it look like for the AI companies? Yeah. Well, I think for authors and publishers, this is hopefully a new revenue model for them.
SPEAKER_01They're getting AI licensing revenue. I think it could be actually a really big revenue source. For the book industry, there was print books, and there was e-books, and there was audio books. We think that AI revenue is the next big revenue source for the industry. I also think that authors and publishers can reach readers in new ways. It might be that people... don't want to read a particular book and it's kind of linear form, but if you can present it to the consumer in a new way, you can actually get more engagement on that, on that book. So I think hopefully we'll be able to connect,
SPEAKER_01um, you know, authors and publishers with audiences in new ways. Um, for AI companies, I'm hoping they can, they can be building, um, you know, next generation products, but doing it in a way where they have total confidence they're doing it legally. Um, and they, they don't have all these lawsuits and all of this, um, you know, all of this, uh, angst from, from the creative community. So I hope it's working, um, um, working that way for them. So I'm hoping we can just kind of make this a win-win for everyone involved.
SPEAKER_00Nice. There's vision of peace between. Yep. Yep. We're peacemakers between the human and the machine. That's what we're trying to do. Yep. Trip, we're going to leave it there. Thanks so much for taking the time and telling us all about what you're doing.
SPEAKER_01Yeah, thank you. It was great, great chatting today. Thanks for listening to my conversation with Trip Adler of Created by Humans.
SPEAKER_00If you enjoyed this episode, I'd appreciate it if you'd just take a minute to leave a rating or review for the Media Co-Pilot in your podcast app. Also, don't forget to subscribe to the Media Co-Pilot newsletter and check out mediacopilot.ai for our classes on AI for journalists, PR professionals, and communicators. Be seeing you in the future.
For years, Yahoo News has occupied a peculiar position in the media ecosystem. It is not a social network, not a traditional publisher, and not a search engine in the modern sense. Yet millions of people still use it as...
This special episode of The Media Copilot is a look back at the conversations that defined the past year… the questions, tensions, and turning points shaping how media, journalism, and technology intersect right now....
AI has made verification a newsroom survival skill. In this episode of The Media Copilot, Pete Pachal talks with James Law, Editor-in-Chief of Storyful, the newsroom partner known for verifying the videos and social...
AI in journalism can feel abstract until you talk to the people actually shipping products inside newsrooms. In this episode of The Media Copilot podcast, host Pete Pachal talks with Darla Cameron, Chief Product Officer...
AI is rewriting the rules of digital media and few people have had a closer view of the shift than Adam Singolda. In this episode of The Media Copilot, host Pete Pachal has a candid conversation with the Taboola founder...